Skip to main content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Analysis of Children's Health Insurance Patterns: Findings from the SIPP

Publication Date
May 11, 1999

May 12, 1999
John L. Czajka

Submitted to:

Department of Health and Human Services
Office of the Secretary
Assistant Secretary for Planning and Evaluation
HHH Building, Room 442E
200 Independence Avenue
Washington, DC 20201

Submitted by:

Mathematica Policy Research, Inc.
600 Maryland Ave., S.W.
Suite 550
Washington, DC 20024
(202) 484-9220

Project Officer:

Rob Stewart
Sarah Schoenecker

Project Director:

John L. Czajka

"

Acknowledgments

Several people contributed to this report and the analyses that underlie it. The author is particularly grateful to Marsha Gold and Margo Rosenbach for their detailed review of the report and the many valuable suggestions that they provided. The author would also like to thank Jim Verdier for his help in highlighting and interpreting key findings, and Embry Howell and Harold Beebout for their overall guidance in the final stages of this project. Finally, the author is grateful to Bob Cohen for his many hours of careful programming and table construction; to Marilyn Ellwood and Kimball Lewis for answering countless questions about data and eligibility rules; to Daryl Hall for her excellent editing; and to Felita Buckner, Pat McCall, and Angie KewalRamani for their good work on the final production of the report.

Executive Summary

This paper summarizes work using the 1992 panel of the Survey of Income and Program Participation (SIPP) to provide the most detailed look yet at the dynamics of health insurance coverage among children and the relationship between Medicaid eligibility and insurance coverage. We summarize what our technical work (contained in companion technical appendices A-D) says on five broad questions:

  1. How many children lack insurance?
  2. How long do children remain uninsured?
  3. What are the patterns of movement between spells with and without insurance?
  4. What are the characteristics of children by type of health insurance coverage?
  5. How many fewer uninsured children would there be if participation in Medicaid were more complete?

We find that from October 1992 to September 1994, between 12.4 and 13.3 percent of children under 19 were uninsured at any one point in time, with one in four children (27.1 percent) uninsured for some period of time over the two years. About half of all spells without insurance last less than six months, but a quarter last 12 months or more. Further, nearly half of all children uninsured at any one time were already uninsured for 12 months or more.

Among children becoming uninsured, 52 percent were covered by employer-sponsored coverage in the previous month, 39 percent by Medicaid, and 9 percent by other sources or were newborns. More than half of the 550,000 children leaving Medicaid each month (56 percent) became uninsured. But loss of Medicaid eligibility explains fewer than half these transitions. While reporting error may explain some of this, our findings highlight the importance of shifting some attention from studying why children do not enroll in Medicaid to the question of why children lose Medicaid coverage when they may still be eligible. This is particularly relevant as our findings highlight the sensitivity of insurance coverage rates to Medicaid enrollment patterns.

The 1992 SIPP data support previous research on the relationship between children's insurance status and various demographic and socio-economic characteristics. The findings highlight how Medicaid eligibility policy affects the relationships among income, child age, and coverage. They also highlight the periodic discrepancy between a child's coverage and that of the parents. About one-fifth of uninsured children appear to have at least one insured parent, although some of this may be reporting errors. Children who experience long spells without insurance do not appear strikingly different from those who experience relatively brief spells, contrary to what some may expect.

Our data suggest that Medicaid participation rates by children range from 65 percent to 79 percent, depending on how they are calculated. Simulations indicate that about one third of the 8.9 million uninsured children in September 1994 were eligible for Medicaid; some of this group is in transition to Medicaid, however, and many more were enrolled previously. The findings also highlight the substantially lower participation rates for children who are not cash assistance participants. This reinforces the importance of strong outreach to achieving high levels of CHIP participation and highlights ways of targeting outreach to reach diverse subsets of children.

In sum, our work suggests that the number of children with some spell without insurance over time greatly exceeds the number uninsured at any point in time. The findings highlight the importance of not just enrolling children in Medicaid or CHIP but also retaining them in the programs. They also show that, contrary to common perception, Medicaid participation by eligible children with no other coverage was relatively high before welfare reform. These rates were lifted by the high proportion of children covered through cash assistance programs, reinforcing the concern over the potentially adverse effects of TANF on Medicaid coverage for children.

Purpose of This Report

The passage of the legislation establishing the Children's Health Insurance Program (CHIP) has turned the policy spotlight on the number and characteristics of children without health insurance. In fact, many more children go without health insurance than the number who are uninsured at any one time. The research reported here was conducted to explore the dynamics of health insurance coverage among children and to tell us more about the relationship between Medicaid eligibility and insurance coverage. The unique contribution of this research is that it provides the most detailed picture yet of the dynamics of health insurance coverage among children by addressing five broad questions:

  1. How many children lack insurance?
  2. How long do children remain uninsured?
  3. What are the patterns of movement between spells with and without insurance?
  4. What are the characteristics of children by type of health insurance coverage?
  5. How many fewer uninsured children would there be if participation in Medicaid were more complete?

This report summarizes the key findings in each of these areas and interprets them against the backdrop of current policy discussions.

Source of Data

This analysis is based on the 1992 panel of the Survey of Income and Program Participation (SIPP), a longitudinal survey that provides extensive monthly data on a large sample of children (and adults) from 1992 through 1994. These are the latest data with which we could construct the multi-year longitudinal measures presented in this report. The 1993 SIPP panel would have given us data through 1995, but our work with these data raised concerns that the number of poor children was overstated. The Census Bureau started another SIPP panel in 1996, but until recently, only the first wave of data, covering a four-month reference period, had been released.

Given that readers will want to apply the findings presented here to the present, an important caveat is that the data on which this analysis is based precede welfare reform. Clearly, point estimates of uninsured children and Medicaid participants in 1994 will differ from today's figures, but the underlying dynamics on which this report focuses remain relevant even though some of the particulars may have changed.

How Many Children Lack Insurance?

In the two years from October 1992 through September 1994, between 12.4 and 13.3 percent of children under 19 were reported to have been uninsured at any one time. Over the course of each fiscal year, however, 21.5 percent of children went at least one full month without insurance, and over the two-year period, 27.1 percent--more than one in four children--were uninsured for some period of time. Policymakers tend to focus on the point-in-time estimates, but figures representing periods of a year or more are at least as important because they tell us the full extent to which children are exposed to spells without insurance. Annual data from the Current Population Survey (CPS) suggest that before the widespread implementation of CHIP this exposure was increasing. Indeed, both the number and percentage of children who were without insurance rose after 1995 (Fronstin 1997a, 1997b).

How Long Do Children Remain Uninsured?

About half of all spells without insurance end in less than six months, but a quarter last for 12 months or more. By examining all spells that started in 1993 and following them to their conclusion or to the end of 1994, whichever came first, we could identify spells that continued for at least 12 months and those that ended in any number of months up to 12. More than 9 million new spells began in 1993 (including multiple spells experienced by the same children). So the number of new spells is comparable to the number of children who were without insurance at any one time. That half of these new spells ended in less than six months helps to explain why the number of children who were uninsured for any part of 1993 is so much higher than the number who were uninsured at any one time during the year. Nevertheless, a significant fraction of the new spells--about 24 percent--did last for 12 months or more. Therefore, new spells that started in 1993 added more than 2.3 million children, ultimately, to the number of long-term uninsured (based on a common 12-month definition of long-term).

Nearly half of the children who are without insurance at any one time have already been uninsured for 12 months or more. Of the more than 9 million children who were uninsured in September 1993, 47 percent had been uninsured for at least 12 months, while only 37 percent had been uninsured for less than 6 months. We projected that fewer than 20 percent of these 9 million children would end their spells (become insured) in less than six months and that nearly two-thirds would be uninsured for 12 months or more. Essentially, long spells are more likely to be ongoing at any one time than short spells, so the uninsured population at a point in time, includes a higher proportion of long-term uninsured than we find among all new spells.

The distribution of spell durations is important because brief spells, moderate spells, and long spells without insurance carry very different policy implications. By definition, very short spells are transitional. That is, with no intervention, these spells will end quickly, the assumption being that these uninsured children are in transition from one source of insurance to another (perhaps because of a waiting period for private insurance). These transitional periods, or short spells, demand a different policy response, if any, than periods in which the near-term availability of coverage may be uncertain. Even spells of six months or so may require a different policy approach than much longer spells--both from the standpoint of program efficiency and potential clients' perceived needs for coverage.

When a spell starts, how do we determine if it is going to be a short spell or a long spell? The surest approach is to wait. The CHIP regulations explicitly address the differential treatment of short-term and long-term spells, and the federal guidelines allow waiting periods as a device for countering the potential crowd-out of private insurance.

What Are the Patterns of Movement Between Spells with and Without Insurance?

What Precedes Spells without Insurance?

Between October 1992 and September 1994, children started nearly 19 million new spells without insurance. Each month, on average, about 800,000 children became uninsured. In the month preceding each new spell, 52 percent of these children were reported to have been covered by employer-sponsored insurance, 39 percent by Medicaid, and 6 percent by another source of insurance. The remaining 3 percent were born uninsured.

Altogether, more than 550,000 children left Medicaid each month, on average. More than half of this group--56 percent--started new spells without insurance. Just over one-third, or 34 percent, gained employer-sponsored coverage, 2 percent obtained coverage from another source, and 7 percent turned 19 or became ineligible for the SIPP.1

These findings beg the question: why do so many children leave Medicaid only to become uninsured? Loss of Medicaid eligibility appears to explain fewer than half of these transitions from Medicaid to uninsured. That is, more than half of the children who left Medicaid and became uninsured still appeared to be eligible for Medicaid, according to a state by state simulation of the Medicaid eligibility rules. It is possible that reporting error may account for some of these apparent departures from Medicaid by eligible children, contributing to the recognized undercount of Medicaid enrollment.2 Research in progress will shed light on the potential role of reporting error in particular, and on the circumstances surrounding these changes in Medicaid enrollment and insurance coverage more generally. Even if the number of children moving from Medicaid to uninsured proves to be fewer than the estimate reported here, policymakers need to shift some attention from the question of why children are not enrolling in Medicaid to the question of why they are losing coverage, especially when many of them may still be eligible.

What Follows Spells without Insurance?

Of the more than 9 million children who were uninsured in any one month, about 8 percent, or 750,000 children on average, became insured in the next month, and another 60,000 simply aged out of the population of children (turned 19) or left the SIPP universe. Medicaid had a disproportionate impact on these transitions. Of the children who became insured each month, 41 percent enrolled in Medicaid, 52 percent obtained coverage from their parents' employers, and 7 percent gained coverage from another type of plan. Thus, while Medicaid was reported to insure only 19 percent of all children under 19 at any one time in 1993 or 1994, it accounted for 41 percent of the uninsured children who gained coverage. This suggests that the number of uninsured children will be particularly sensitive to changes in Medicaid enrollment patterns. Indeed, the decline in the Medicaid caseload that followed the introduction of welfare reform appeared to produce a full percentage-point increase in the proportion of children who were without health insurance, as measured in the CPS (Fronstin, 1997b).

How Much Churning Is There among Uninsured Children and Medicaid Participants?

"Churning" in a population consists of exits and re-entries by the same group of people over a relatively short period, whereas "turnover" more literally involves exits and entries by different people. The frequency of churning and turnover among uninsured children and Medicaid participants influences the magnitude of the task of insuring the uninsured and the efficiency with which it may be possible to do so. For example, churning in the Medicaid population has come to be viewed as a serious problem by Medicaid managed care organizations.

As we noted, 12 million children started nearly 19 million new spells without insurance between October 1992 and September 1994. In other words, more than one-third of the new spells that began during this two-year period were the second or third spells of children who had become uninsured earlier in the period. In each year alone, 20 to 23 percent of the new spells were additional spells of children becoming uninsured for at least the second time in the year. The magnitude of this churning among the uninsured raises questions that must be addressed if we are to make significant progress in permanently reducing the number of uninsured children. Why do so many children regain coverage only to lose it again within a year or two? What are the roles of public and private sources of insurance in accounting for this phenomenon?

If we look at who is uninsured from one year to the next, we find less turnover than we might imagine, given the sizable numbers of children entering and leaving the uninsured population each month. While 800,000 children regained insurance coverage each month over the study period, only to be replaced by another 800,000 who lost coverage, and while 12 million different children became uninsured over the two-year period, we find that 62 percent of the 9.5 million children who were uninsured in October 1992 were still uninsured in September 1993, and 52 percent were still uninsured a year later. In other words, after two years, only about half of the uninsured children or between four and five million had been replaced by children who became uninsured during the intervening period. Some of the children who were still uninsured had undoubtedly remained uninsured continuously over the two-year period, but the estimates of churning suggest that children who gained insurance coverage and then lost it again may account for a significant proportion of the children who were uninsured at both the beginning and end of the period.

While the turnover in the population of uninsured children over one or two years may appear small, we would underscore two aspects of these findings:

  • The turnover is not small in terms of actual numbers of children. If the federal government and the states could immediately and permanently insure the estimated 11 million children who are currently uninsured, these findings suggest that a year later, an additional 4 to 5 million children would be uninsured.
  • The figures on year-to-year turnover hide the many children who go without insurance for some period during the intervening 12 months--losing but regaining coverage before the end of the year. Again, in the scenario that we described, with current estimates of the uninsured, more than 4 million additional children would go without insurance for some period during the year.

Relative to the number of new enrollments, the amount of churning in the Medicaid program is on a par with what we see among uninsured children. Specifically, 9.0 million children initiated the 13.4 million new spells of Medicaid participation in FY93 or FY94. That is, one-third of the new spells were second or third spells by children who had enrolled in Medicaid earlier in the period. Each year, about one in five new spells included children who were enrolling in Medicaid for at least the second time during the year.

Both in absolute number and as a proportion of the total caseload, however, new spells on Medicaid were less common than new spells without insurance. The 13.4 million new spells on Medicaid that started between October 1992 and September 1994 contrast with the 19 million new spells without insurance initiated over the same period. The month-to-month turnover in the Medicaid child population, at 4 percent of the caseload, was only half the monthly turnover in the uninsured. Further, 80 percent of the children who were enrolled in Medicaid in October 1992 were still enrolled in Medicaid in September 1993 and, a year later, 76 percent. Again, some portion of the children who were enrolled in Medicaid at the beginning and end of the period actually left the program and returned, but given the lower exit rates from Medicaid than from the uninsured, we would expect a larger share to have been continuously enrolled in Medicaid than the fraction of uninsured children who were continuously uninsured over the same period.

In sum, while churning has attracted interest in studies of the Medicaid population, we find more churning and substantially more turnover among uninsured children than among children enrolled in Medicaid. However, we must recognize that changes in the Medicaid caseload that are developing post-welfare reform could begin to alter the Medicaid dynamics as well.

Notes

1. Children leave the SIPP population by moving out of the country, becoming institutionalized, moving into military barracks, or dying.

2. Surveys report fewer children enrolled in Medicaid than the statistics compiled by the Health Care Financing Administration (HCFA(now known as CMS)) indicate. This finding is typical of comparisons between survey data and program administrative statistics for means-tested entitlement programs. Some researchers adjust their estimates of uninsured children to compensate for Medicaid underreporting (Ullman et al. 1998). Others caution that we know too little about the sources of the Medicaid undercount to reliably attribute a large portion of it to children who are misreported as uninsured. In our case, adjusting reported insurance coverage to compensate for a Medicaid undercount was not practical, for the most part, because there is no evidence on how Medicaid underreporting might be correlated with the complex dynamics and the many characteristics we examined. Nevertheless, it is important to recognize when we and other researchers examine Medicaid eligibility among uninsured children that the survey data on which we base our estimates fail to account for all children who are counted in the Medicaid program statistics.

What Are the Characteristics of Children by Health Insurance Coverage?

How Does Insurance Coverage Vary among Subgroups of Children?

Other research has established that uninsured children differ from insured children in terms of a variety of demographic and economic characteristics (Fronstin 1997b, Lewis et al. 1997). In our analysis, insurance coverage varied with nearly every characteristic that we examined and in ways that largely replicated earlier findings. For example, Hispanic children were more than twice as likely to be uninsured as black children, who were only marginally more likely to be uninsured than white children. Nearly three-quarters of children whose parents did not work were enrolled in Medicaid compared to about half that percentage among children whose parents were employed part-time. This latter group along with children with no parent in the household were the most likely to be uninsured, but together, they accounted for only 13 percent of all uninsured children.

While employer-sponsored insurance was the most common source of coverage, subgroups of children with comparatively low rates of employer-sponsored coverage were not necessarily more likely to be uninsured. Medicaid accounts for this situation but fills the gap more completely for some groups than others. Children in families below 50 percent of poverty, for example, have only one-fifth the employer-sponsored coverage that children between 100 and 200 percent of poverty have, but they are less likely to be uninsured. Indeed, uninsured children in families between 100 and 200 percent of poverty account for nearly 40 percent of all children without insurance, while those below 100 percent of poverty account for one-third of all children without insurance. This pattern is evidence of the need for the broad coverage that states are allowed to provide through CHIP.

How Does a Child's Insurance Coverage Compare to That of His or Her Parents?

Children's coverage tends to mirror that of their parents, but the exceptions to this rule pose interesting questions for policymakers and analysts. About one-fifth of uninsured children appear to have at least one insured parent--a situation that merits further research to establish if and why it is true. Between 7 and 8 percent of uninsured children, for example, have a parent who reports being covered by Medicaid--a fraction that appears much too high and suggests that many of these children may in fact be covered by Medicaid, contributing to the aforementioned Medicaid undercount . At the same time, between 10 and 13 percent of the children who were reported to have been covered by Medicaid in the study period appeared to have an uninsured parent. This finding is very plausible in light of the opportunity for many children to be enrolled in Medicaid under child-only provisions. In fact, we would very likely find an even greater proportion of Medicaid children with uninsured parents if the Medicaid participation rates of children eligible under the various child-only provisions were as high as they are in the eligibility categories with the greatest participation. This situation raises the policy question--to which we will return--of whether the Medicaid enrollment of children could be increased by extending eligibility to parents as well.

How Does a Child's Age Influence the Chances of Being Uninsured or on Medicaid?

It is important to measure the relationship between age and health insurance coverage because age differences in coverage reflect experiences over the life cycle. Both the probability of being uninsured and the average time without insurance increase with age--a finding established by longitudinal analysis. To generate findings like this, we followed children over the entire year at a given age. Between infancy and late adolescence, the probability of being uninsured during a year rose from 16 percent to 27 percent. Among those who were uninsured for any length of time, only 10 percent of infants versus 38 percent of 18-year-olds were uninsured for the entire year.

The likelihood of being both uninsured and ineligible for Medicaid increases even more sharply with age than does being uninsured alone. Only 6 percent of infants compared with a full 24 percent of 18-year-olds were ever uninsured and ineligible for Medicaid during the year. Furthermore, while only 5 percent of those infants who were ever uninsured and Medicaid-ineligible remained so for the entire year, nearly one-third of the 18-year-olds in the same situation remained so for the entire year.

Medicaid coverage declines sharply with age. Age differences in Medicaid participation are to a large degree a reflection of eligibility criteria that progressively favor younger children.1 Our findings highlight the very prominent role of Medicaid among infants and how that role diminishes as children grow older. Nearly 40 percent of all infants were covered by Medicaid for some part of their infancy, and nearly two-thirds of them, or 25 percent of all infants, were covered for the entire year. With increasing age, the probability that a child was ever covered by Medicaid during the year declined gradually to about 15 percent or less, while the proportion of Medicaid children who were covered for the entire year fell to about one-third. In contrast to the 25 percent of all infants who were covered by Medicaid for the entire year, only 10 percent of 10-year-olds and less than 5 percent of 18-year-olds were enrolled for the entire year. This latter differential is likely to diminish over time as Medicaid eligibility under the poverty-related criteria is phased in and Medicaid expansions are introduced under CHIP.

What Characteristics Distinguish Uninsured Children Who Are Medicaid-Eligible?

In general, uninsured children who are Medicaid-eligible resemble Medicaid participants, which supports the contention that many of these children are either in transition to Medicaid or are enrolled but have not reported their coverage. However, these two groups of people differ notably from each other in four areas:

  • Racial Composition. Uninsured Medicaid-eligible children are twice as likely to be white as black, whereas the proportions among Medicaid participants are nearly equal. Hispanic children are only slightly more represented among the uninsured Medicaid-eligibles than among Medicaid participants.
  • Family Structure. Children from two-parent families account for more than half of the Medicaid-eligible uninsured whereas they are one-third of Medicaid participants.
  • Poverty Status. Children in families between 100 and 200 percent of poverty make up one-third of the Medicaid-eligible uninsured compared to one-fourth of Medicaid participants.
  • Parents' Employment Status. Children with a parent employed full-time make up half of the Medicaid-eligible uninsured but only 30 percent of Medicaid participants. The reverse is true of children whose parents are not working.

These findings have direct implications for Medicaid outreach. For instance, the characteristics of eligible children who are less likely to participate can guide decisions about how--and where--to target the outreach and education efforts under both Medicaid and CHIP to achieve higher participation. But what accounts for the lower participation probabilities of these groups of eligible children? Further research to answer this question could significantly enhance the design of outreach programs.

Do the Characteristics of Uninsured Children Vary by Spell Duration?

Differences in the characteristics of children with long versus short spells without insurance are also potentially important to the design of programs as well as targeted outreach. However, children who experience long spells without insurance do not appear to be strikingly different from children who experience relatively brief spells without insurance. Researchers have found very strong differences between children with and without insurance coverage, but once children lose their coverage, there appears to be little that differentiates between those who remain uninsured for long periods of time and those who become re-insured fairly quickly. Children with long spells were somewhat more likely to be Hispanic and to live with both parents, but no other notable differences emerged. On the whole, these findings give us little information to use in designing programs or outreach.

Notes

1. For example, under provisions included in the Omnibus Budget Reconciliation Act, states were required to cover infants and children under age 6 in families with incomes below 133 percent of the federal poverty level and to cover older children born after September 30, 1983 in families below 100 percent of poverty. Many states exercised an option to cover infants up to higher income levels, and a smaller number extended the income limits for older children as well. In addition, families with younger children were more likely than families with older children to have an income below the levels that qualified them for Aid to Families with Dependent Children (AFDC) and thus were more likely to obtain Medicaid coverage through that route.

How Many Fewer Uninsured Children Would There Be with More Complete Participation in Medicaid?

How High Is Medicaid Participation?

With no adjustment for the Medicaid undercount in the SIPP, the reported participation in Medicaid is 65 percent among children who we simulated to be eligible.1 If we exclude from the denominator those eligible nonparticipants who reported some other source of coverage, the participation rate rises to 79 percent. With an undercount adjustment, the 65 percent figure would rise to about 75 percent.

What Do Medicaid Participation Rates Tell Us about Strategies for Outreach?

Medicaid participation rates among children vary widely by the basis of eligibility. In September 1994, we observed the following participation rates:

  • 100 percent among AFDC participants, who are automatically enrolled
  • 75 percent among reported SSI recipients and foster children
  • 54 percent among children in families that appeared to be eligible for AFDC but did not report that they were receiving cash assistance
  • 37 percent among children eligible under the poverty-related expansions
  • 23 percent among other eligibility groups3

As the number of families receiving cash assistance diminishes under welfare reform, we would expect the overall Medicaid participation rate to drop significantly unless there is a sizable increase in participation rates among children in these other eligibility categories.

The overall pattern of participation and the nature of changes in the cash assistance caseload post-welfare reform carry implications for CHIP. First, it is clear that outreach efforts will have to be very strong to achieve high participation rates in both the Medicaid expansions and the state programs. In the absence of strong outreach, we might expect to see participation rates that are no better than what we observed in the populations eligible for Medicaid without cash assistance. Second, to the extent that CHIP extends eligibility to families and children who were not eligible previously, participation rates could be even lower than we have seen for non-cash families and their children historically. Many newly eligible families may not perceive that a public program is truly for them; they may not even be aware that they are eligible and, lacking experience with public programs, may not think to inquire. They may also hesitate because of the stigma that they believe accompanies participation in such programs. Third, different outreach strategies may prove to be differentially effective with different eligibility groups and different subgroups of the population. At a minimum, strategies should be designed to deal with three different types of eligibles, who will be found in different proportions in different eligibility groups and population subgroups:

  • Children who have been eligible for Medicaid but resisted participation
  • Children who have already participated
  • Children who are becoming eligible for the first time

It may be relatively easy to enroll children who have already participated, but it is likely to be much more difficult to enroll children who have resisted participation or who have become eligible for the first time. For the newly eligible, outreach will have to communicate basic information about the program and allay the concerns of prospective new clients. For those who have been eligible but have not participated, we may need to understand the reasons for their behavior in order to design effective outreach programs.

Another factor that plays into this issue of outreach is the role of coverage for the parent. The differences in participation by eligibility group provide at least a suggestion that children are more likely to participate if a parent is eligible as well. Some states have begun to consider strategies that would allow them to use CHIP funds to partially fund coverage for the parents of eligible children. Further research into the importance of parents' access to coverage in determining their children's likelihood of participating in public insurance programs is clearly in order and may lend additional support to efforts intended to increase the coverage that is available to parents.

What Proportion of Uninsured Children Are Eligible for Medicaid?

Our simulation of Medicaid eligibility revealed that 2.9 million, or 33 percent, of the estimated 8.9 million uninsured children in September 1994 were eligible for Medicaid. This is comparable to what other researchers have reported with more limited Medicaid eligibility simulations (Lewis et al. 1997).2 With longitudinal data, however, we could also address the question of what happens to these Medicaid-eligible uninsured children over time; and what we found suggests that policymakers need to think about the Medicaid-eligible uninsured differently than they may have in the past.

Our findings do not support the inference that all 3 million of these children would be enrolled in Medicaid if only Medicaid outreach were more effective. First, much of this group is in transition--including transition to Medicaid. Between one month and the next about 20 percent of the Medicaid-eligible uninsured children completed the transition, with roughly equal shares enrolling in Medicaid, obtaining employer-sponsored or other coverage, or experiencing a rise in family income that made them ineligible for Medicaid. Within five months, more half of the initial 2.9 million Medicaid-eligible uninsured children had left that state. The mobility among the Medicaid-eligible uninsured suggests that while all of these children could theoretically be brought into Medicaid, or brought into Medicaid more quickly, it is useful to distinguish between those who would remain eligible and otherwise uninsured for only a few months and those who would remain in that state for a much longer period. Second, 25 percent of the children who transition in and out of Medicaid eligibility and between 40 and 50 percent of those who remain Medicaid-eligible and uninsured for longer periods were actually enrolled in Medicaid previously. Thus, there is a sizable group of uninsured and Medicaid-eligible children for whom it may be more relevant to ask why have they left Medicaid than why have they not enrolled. Better Medicaid outreach is not the answer to removing them from the uninsured.

In 1994, about one million children remained uninsured for a year or more despite being eligible for Medicaid, and the number is likely to have grown since then. Policymakers would do well to focus on this group, whose persistence in this condition suggests that they are hard to reach and so will require extraordinary outreach techniques if they are to be brought into Medicaid. The demographic and socioeconomic characteristics that we examined shed little light on what differentiates this group of children from those who more quickly become insured or Medicaid-ineligible. Ultimately, understanding why children remain uninsured and Medicaid-eligible for a year or more may require different kinds of data than our surveys can provide, but we have not exhausted what longitudinal surveys like the SIPP can tell us.

Notes

1. Our simulation did not encompass all categories of eligibility; nor could it take account of the full range of income disregards that states might apply to individual cases. For this reason, we do not include in our participation rate those children who were reported as Medicaid participants but were simulated to be ineligible.

2. While the level of detail in our simulations and the use of monthly income would tend to raise estimates of eligibility relative to many of these other efforts, the SIPP obtains more complete reporting of Medicaid participation than most other surveys, which reduces the number of simulated eligibles who are identified as uninsured.

3. This category does not include children whose eligibility depended on their families "spending down" their income to meet thresholds specified in a state medically needy program. We did not simulate this aspect of Medicaid eligibility because data on health care expenditures are very limited in the SIPP.

Conclusion

Examining the dynamics of health insurance coverage among children tells us much about the magnitude of the task of insuring the uninsured. While policymakers and researchers tend to focus on the number of children who are without insurance at any one time, children who experience one or more months without insurance over the course of a two-year period are more than double this number. Most children who become uninsured remain uninsured for less than six months at a time, so growth in the number of uninsured children has remained modest. Nevertheless, policymakers' well-warranted concerns about the difficulty of enrolling a high proportion of eligible children--especially those who have not participated in Medicaid previously--bring the problem into focus. Turnover in the population of uninsured children implies that outreach efforts will have to continue at a high level in order to achieve and maintain strong participation in Medicaid and CHIP.

Retention of children who have already enrolled in Medicaid or CHIP emerges as a potentially important policy issue. More than half of the children who leave Medicaid are uninsured the next month, and such children account for two-fifths of the newly uninsured. Why do they leave Medicaid, and if access to insurance continues to be a need, what can be done to keep them enrolled?

Despite a widespread perception to the contrary, participation in Medicaid by eligible children with no other insurance was quite high before welfare reform, and many of those who appeared to be uninsured and eligible for Medicaid were actually in transition--often to Medicaid. But participation rates were lifted by the high proportion of cash assistance recipients. As this group declines in importance the historically much lower participation rates of other groups of eligibles may become more typical. As CHIP extends affordable health insurance coverage to increasingly more children, the greatest challenge for policymakers is to ensure that no child who needs this coverage is denied it.

References

Fronstin, Paul. "Trends in Health Insurance Coverage." EBRI Issue Brief no. 185. Washington, DC: EBRI, May 1997a.

Fronstin, Paul. "Sources of Health Insurance and Characteristics of the Uninsured: Analysis of the March 1997 Current Population Survey." EBRI Issue Brief no. 192. Washington, DC: EBRI, December 1997b.

Lewis, Kimball, Marilyn Ellwood, and John L. Czajka. "Children's Health Insurance Patterns: A Review of the Literature." Washington, DC: Mathematica Policy Research, Inc., 1997.

Ullman, Frank, Brian Bruen, and John Holahan. "The State Children's Health Insurance Program: A Look at the Numbers." Assessing the New Federalism, Occasional Paper No. 4. Washington: The Urban Institute, 1998.

Technical Appendix A: Measuring Medicaid Eligibility, Medicaid Participation, and Lack of Health Insurance Among Children: Exploring the Issues with the SIPP.

Using data from the Survey of Income and Program Participation (SIPP), this report examines issues related to the measurement of health insurance coverage and Medicaid participation and the simulation of Medicaid eligibility and draws some observations about the sources of health insurance coverage among children, the frequency with which children lack health insurance coverage, and the frequency with which these same children are eligible but apparently not enrolled in Medicaid. Subsequent reports will investigate the dynamics of health insurance coverage and Medicaid eligibility and participation among children and the characteristics of children by their health insurance coverage.

This report is organized as follows. Section A provides an overview of the SIPP, including its strengths and limitations, its general representativeness, and its representation of the population of children over time. Section B discusses our use of the SIPP to measure children’s health insurance coverage. Section C presents estimates of health insurance coverage among children. Section D details our methodology for simulating Medicaid eligibility and presents estimates of Medicaid- eligible children by their basis of eligibility. Section E presents estimates of Medicaid participation and other insurance coverage among Medicaid-eligible children, and Section F discusses strategies for evaluating the Medicaid eligibility simulation and presents some comparisons with program administrative statistics as well as estimates of the impact of one feature of our simulation--the definition of the family unit.

A. An Overview of the Sipp

The data on which this report is based are from the 1992 panel of the SIPP. The SIPP is a longitudinal survey whose respondents are interviewed every four months about their activity during the preceding four months. The questions include a lengthy series of “core” items, included in every interview, and periodic “topical modules” that collect data more infrequently on specialized areas. One quarter of the sample, constituting a “rotation group,” is interviewed in every month, so that the data for a given calendar month are based on a roughly equal distribution of respondents answering questions about activities one month ago, two months ago, three months ago, and four months ago. The staggered interviewing is intended to ensure that no calendar month of data is affected unduly by recall bias or other error associated with distance from the interview.

The Census Bureau collected nine waves of data--that is, nine interviews--from the entire 1992 SIPP panel sample and a tenth wave of data from three of the four rotation groups (that is, three- quarters of the sample). These data provide a common reference period covering three full calendar years--1992, 1993, and 1994--although, as we will explain, the Census Bureau is not releasing all of the data collected for the final three months of 1994.

1. Strengths and Limitations of SIPP

Several features of the SIPP make these data especially appealing for the analysis of children’s health insurance coverage. The SIPP provides a detailed measure of health insurance coverage for every month of the two to three year duration of a panel. Because of the SIPP rotation group design, the estimates for a given calendar month are based on a median two-and-a-half-month recall, with one quarter being only one month and one quarter being four months. Measures of the duration of new spells of particular types of coverage or lack of coverage can be constructed by aggregating the reports from successive interview waves, so that no matter how long the measured duration of a particular spell, no part of a reported spell relies on respondent recall beyond four months. In addition to providing measures of health insurance coverage, SIPP also provides very detailed measures of demographic and economic characteristics--again on a monthly basis. This affords us the opportunity to construct contemporaneous measures of circumstances that may affect eligibility for and enrollment in particular types of coverage.

Despite the strengths of the SIPP design, however, there are some notable limitations. Monthly reports of a number of characteristics--including Medicaid participation and uninsurance--show evidence of a pronounced “seam effect.” That is, monthly transitions (for example, in reported Medicaid coverage or health insurance coverage generally) are reported as occurring disproportionately between the four-month reference periods of interview waves rather than within these reference periods. Whereas we would expect only one quarter of such transitions to occur between reference periods, we find evidence that 75 to 90 percent of the transitions in certain statuses occur between interview waves--as if respondents were reporting their coverage (or interviewers recording them) in four-month chunks rather than month-by-month. As we demonstrate in another report [no citation yet], this has a profound impact on the reported distribution of spells of Medicaid coverage and uninsurance, and it may affect the point-in-time estimates as well (unless respondents are equally likely to “round up” as “round down” their reported months of coverage in a reference period). Like other surveys, SIPP shows evidence of underreporting of program participation and many sources of income, although there is evidence to suggest that SIPP does better in this regard than surveys with annual reference periods and less detailed measurement of these characteristics.

2. Representativeness of the SIPP Panel

It is important to recognize the implications of the SIPP’s longitudinal design for the representativeness of the information that it collects. The SIPP sample is selected from the resident population of the United States, excluding persons living in military barracks or institutions. The sample is designed to be representative of this population at the time that it is drawn, and the initial respondents are weighted to Census Bureau estimates of the size of this population by age, sex, race, and Hispanic origin. The SIPP sample is dynamic, however. Over the life of a SIPP panel some respondents leave the sample and others are added. As a result, the sample size and its representativeness change over time.

Attrition. Respondents who refuse to continue participating in the survey, move to an unknown address, or move more than 100 miles from a SIPP primary sampling unit and cannot be interviewed by telephone are lost from the panel. Because they continue to belong to the population that the SIPP panel was selected to represent, the sample weights of other sample members must be adjusted to compensate for their loss. The 1992 panel had a 9.3 percent nonresponse rate to the initial interview and a cumulative nonresponse rate of 26.2 percent through the ninth interview.

Exits from the Population. Persons who die, move outside the country, enter institutions, or move into military barracks leave the population as well as the sample. Because these losses affect the population as well as the sample, they are not treated as attrition. There is no adjustment to the weights of other panel members to compensate for their loss.

Additions to the Sample. Persons who move into the households of panel members (including those who are born to panel members) become sample members and remain so for as long as they continue to reside with original panel members. Likewise, persons into whose household an adult SIPP panel member moves become sample members as well--again, for as along as the panel member continues to reside with them.

Births clearly represent additions to the population as well as the sample. Other persons added to the sample after the initial interview may or may not represent additions to the population that the SIPP sample represents. Persons who move into or return to the country, leave institutions, or move out of military barracks constitute additions to the population. If they move into SIPP households they become SIPP sample members, and through their addition to the sample the SIPP can be said to represent all persons who joined the population and moved into households that were included in the initial population. Persons who move into the population but form their own households cannot join the SIPP sample. Strictly speaking, then, the SIPP sample does not represent these additions to the population over time. But the SIPP sample weights, as we shall explain, take account of these additions, and so they are represented in number if not actual sample members.

SIPP Weights. To enable inferences from the SIPP sample to the total population the Census Bureau constructs both cross-sectional and longitudinal weights. The cross-sectional weights are created for each calendar month. Weights for a given month are assigned to all persons for whom data were collected in that month, and they are constructed so that they sum to an estimate of the total population by age, sex, race, and Hispanic origin in that month. These weights account for sample attrition (see below) as well as net additions to the population.

The Census Bureau assigns longitudinal weights to all initial sample members who remain through the final interview or leave the survey universe, providing that they miss no more than one consecutive interview.(1) For the 1992 panel, these persons constituted about 74 percent of the initial sample (where the latter includes first wave nonrespondents). These longitudinal weights are adjusted to compensate for persons who attrited through nonresponse (or were never interviewed), and at the outset they sum to the Census Bureau’s estimate of the SIPP population in March 1992. Because of panel members who exit the population as well as the sample, the weighted sample total declines over time. At any point after the first interview, the longitudinally-weighted SIPP panel represents the survivors of the population that the panel represented fully at the start.

The Census Bureau does not assign longitudinal weights to children born after the first interview. These children cannot be weighted with the same scheme that is used for sample members who were actually present for the initial interview, and the Bureau has elected not to apply an alternative weighting scheme.(2) After the first year, then, the weighted longitudinal sample contains no infants. A year later it contains no children under age two, and a year after that it contains no children under age three. For many research purposes--including ours-- this is not acceptable. Therefore, we have followed what has become a commonly used practice of assigning newborns the weights of their mothers.(3)

Adjustments for Nonresponse. Both the cross-sectional calendar month weights and the longitudinal weights take into account characteristics of the panel members who attrited, but the limitations of this adjustment must be recognized. The nonresponse adjustments cannot fully account for the ways in which the attriters may differ from panel members who remain in the sample because some of these differences cannot be known. For example, some of the attrition may be influenced by important changes in circumstances--loss of employment, divorce, birth of a child-- that occurred after the attriter’s last interview. In addition, attriters may simply be different in ways that are not observed but which affect their behavior post-attrition.

3. Representation of Children Over Time

Table 1 presents comparative estimates of the population of children represented by the SIPP panel sample and the population of children represented by the individual calendar month samples. The latter estimates were obtained by summing the calendar month weights by single year of age for selected months. These population totals represent, approximately, the Census Bureau’s estimates of the population that would have been eligible for selection into the SIPP sample in each of the individual months. Estimates of the populations represented by the panel and cross-section samples are compared at four points in time: January 1992, October 1992, September 1993, and September 1994. January 1992 is the common reference month for the four rotation groups in the first wave of the 1992 panel. The next three months represent the beginning, middle, and end of the two-year period defined by FY93 and FY94, or the period on which our analysis is focused.

The first thing to note in this table is that the population of children to which the SIPP panel “weights up” in January 1992 actually exceeds the size of the population implied by the calendar month weights--by about 1.1 million children. We have no explanation for a difference of this magnitude. While the SIPP panel is weighted to estimates of the relevant population in March rather than January and, therefore, would not be expected to reproduce the January 1992 population totals, neither would we expect it to exceed the January 1992 population counts, much less by such a large margin. With this discrepancy in 1992, and the opposite trends in the two series, the two estimates of children under 19 cross between January and October, 1992.

From January 1992 through September 1994, the SIPP panel estimate of the population of children declines by about 2.2 million while the population implied by the calendar month weights rises by 2.9 million. The decline in the SIPP estimates can be attributed to the SIPP panel’s underrepresentation of births, which propagates through the younger ages. In January 1992 the SIPP panel sample represents an estimated 4.5 million infants. By October of that year the number of infants has dropped by nearly 1.4 million to 3.1 million. The number rises some by September 1993 but then drops by half a million by September 1994. Because the children born into the panel in 1992 become the panel’s one-year-olds in 1993 and two-year-olds in 1994, the effect of the underrepresentation of births is compounded. By September 1994 the SIPP panel represents 3.5 million fewer children under the age of three than it does in January 1992. The Census Bureau’s population estimates reflected in the calendar month weights indicate that the size of this population did decline over this period, but by only 150,000.

TABLE 1: COMPARISON OF THE POPULATION OF CHILDREN UNDER 19 REPRESENTED BY THE 1992 SIPP PANELAND THE SIPP CROSS-SECTION SAMPLE: JANUARY 1992 TO SEPTEMBER 1994
  Estimates from the 1992 SIPP Panel Sample Estimates from the 1992 SIPPCross-section Sample
Age Jan 1992 Oct 1992 Sept 1993 Sept 1994 Jan 1992 Oct 1992 Sept 1993 Sept 1994
Total 72,103,000 71,549,000 70,868,000 69,935,000 71,016,000 72,374,000 73,137,000 73,957,000
0 4,490,000 3,123,000 3,469,000 2,943,000 4,214,000 4,058,000 4,082,000 3,984,000
1 4,236,000 4,569,000 3,008,000 3,392,000 4,048,000 4,168,000 4,016,000 4,127,000
2 4,109,000 4,084,000 4,558,000 2,992,000 4,010,000 4,173,000 4,102,000 4,010,000
3 3,969,000 4,006,000 4,043,000 4,501,000 3,910,000 3,964,000 4,192,000 4,176,000
4 4,033,000 3,880,000 4,004,000 4,021,000 3,899,000 3,886,000 3,962,000 4,152,000
5 3,629,000 4,054,000 3,867,000 3,975,000 3,871,000 3,934,000 4,003,000 3,899,000
6 3,854,000 3,598,000 3,993,000 3,854,000 3,795,000 3,790,000 3,876,000 4,068,000
7 3,768,000 3,759,000 3,607,000 3,988,000 3,757,000 3,852,000 3,886,000 3,961,000
8 3,583,000 3,759,000 3,841,000 3,581,000 3,577,000 3,562,000 3,805,000 3,683,000
9 3,814,000 3,641,000 3,532,000 3,837,000 3,912,000 3,888,000 3,787,000 4,015,000
10 3,816,000 3,841,000 3,785,000 3,519,000 3,814,000 3,968,000 3,722,000 3,774,000
11 4,058,000 3,838,000 3,742,000 3,785,000 3,893,000 3,782,000 4,009,000 3,927,000
12 3,584,000 4,045,000 3,931,000 3,736,000 3,616,000 3,863,000 3,723,000 3,793,000
13 3,471,000 3,458,000 3,906,000 3,919,000 3,507,000 3,576,000 3,841,000 3,861,000
14 3,534,000 3,477,000 3,461,000 3,886,000 3,542,000 3,540,000 3,680,000 3,883,000
15 3,748,000 3,449,000 3,376,000 3,443,000 3,517,000 3,628,000 3,620,000 3,713,000
16 3,778,000 3,776,000 3,578,000 3,370,000 3,577,000 3,631,000 3,707,000 3,617,000
17 3,532,000 3,612,000 3,678,000 3,551,000 3,331,000 3,454,000 3,515,000 3,790,000
18 3,095,000 3,580,000 3,490,000 3,642,000 3,226,000 3,656,000 3,609,000 3,523,000
Change from Jan 1992
Total   -554,000 -1,235,000 -2,168,000   1,358,000 2,121,000 2,941,000
0 to 2   -1,060,000 -1,800,000 -3,508,000   127,000 -72,000 -151,000
3 to 18   505,000 565,000 1,341,000   1,231,000 2,193,000 3,092,000
3 to 5   309,000 283,000 866,000   104,000 478,000 548,000
6 to 10   -238,000 -78,000 -56,000   206,000 221,000 647,000
11 to 15   -128,000 21,000 376,000   314,000 796,000 1,100,000
16 to 18   561,000 340,000 156,000   608,000 698,000 797,000

The net difference of 3.35 million between the two samples accounts for most of the 4.0 million children that would be eligible for SIPP in September 1994 but are not represented by the 1992 panel. The remaining .65 million is spread over the ages 3 through 18. The population estimates show this population growing by 3.1 million between January 1992 and September 1994 whereas the SIPP panel shows growth of 1.3 million. The population of children 3 through 18 grows in SIPP because the number of children who move into this age group from younger ages exceeds the number who “age out” at the upper end or leave the population through death, migration, or institutionalization. The SIPP panel estimates of children 3 to 5 grow by more than the Census Bureau’s population estimates because the SIPP panel overrepresents infants in January 1992. In the 6 to 10 age group, the SIPP panel declines slightly over time while the Census Bureau’s population estimates grow by nearly 650,000. In the 11 to 15 age group, the SIPP panel estimate increases by 376,000 while the population estimate rises by 1.1 million. Finally, in the 16 to 18 age group, the population represented by the SIPP panel grows by 156,000 compared to 797,000 for the cross-section sample..

That the differences between the growth trajectories of the SIPP panel and the total population (of SIPP-eligible children) are relatively small over most of the age range suggests that we can generalize from the SIPP panel to the full population fairly readily. It is only at the lower end of the age distribution that we need to be conscious of major differences between the sample and the population. Indeed, with the SIPP panel representing just over three-quarters of the estimated number of children under three in the population, we should be aware of the potential impact on the distribution of characteristics that differ substantially between very young children and older children.

B. Measuring Children’s Health Insurance Coverage

The SIPP ascertains health insurance coverage by means of a series of questions that ask about specific types of coverage or about coverage in general. These questions are placed relatively early in the SIPP interview, when respondents are likely to be more attentive. By contrast, the health insurance questions in the March supplement to the Current Population Survey (CPS) are placed near the end of the interview.

The applicable questions are reproduced below.

20a.(If ... is 65 years of age or older or ... has a work disability) Medicare is a health insurance program for disabled persons and persons 65 or older. People covered by Medicare have a card that looks like this (SHOW FLASHCARD L). Was ... covered by Medicare?

23a.(If ... is 18 or older or the designated parent or guardian of children under 18 years old who live in the household) During the 4-month period, was ... covered by (use local name for Medicaid) or another public assistance program that pays for medical care?

b.May I see ...’s (use local name for Medicaid) card to record the claim number?

23c.(If ... is the designated parent or guardian of children under 18 years old who live in this household) Were any of ...’s children (under 18) covered by (use local name for Medicaid)?

d.(If yes) Which children were covered?

23e.(If 23a or 23c is marked yes) Was (.../(and)...’s children) covered during the entire 4- month period?

f.(If no) In which months was (.../(and)...’s children) covered?

24a.Was ... covered by a health insurance plan at any time during the past 4 months? (Include CHAMPUS, CHAMPVA, and military coverage. Exclude Medicaid, Medicare, and plans paying benefits only for accidents or specific diseases.) (If no, skip to 24k.)

b.Was ... covered by a health insurance plan during the entire 4-month period?

c.(If no) In which months was ... covered?

d.Was ...’s health insurance coverage from a plan in ...’s own name (primary policy holder) or was ... covered as a family member on someone else’s plan? (If own name, skip to 24f.)

24e.Whose plan covered ...? (Skip to 24k.)

f.Was ...’s policy obtained through ...’s current employer or union, through a former employer, through the CHAMPUS or CHAMPVA programs, or in some other way?

g.(If current employer or union or former employer) Did ...’s employer or union (former employer) pay all, part or none of the premium (cost) of this plan?

h.Was ...’s plan an individual plan or a family plan? (If individual, skip to 24k.)

i.Other than ..., which persons in this household were covered by ...’s plan? (Include children as well as adults.)

j.Did ...’s plan cover anyone who did not live in this household during the past 4 months? (If yes) Who did the plan cover?

24k.(If ... is the designated parent or guardian of children under 15 years old who live in the household) Were all of ...’s children under 15 years old covered by a health insurance plan? (Include CHAMPUS, CHAMPVA, and military plans.) (Exclude Medicare, Medicaid, and plans paying benefits only for accidents or specific diseases.)

l.(If no) Which children were covered by a health insurance plan?

24m.(If 24k is yes or one or more children is listed in response to 24l) Were any of these children covered by the plan of someone who did not live in the household during the past 4 months? (If yes) Which children?

These questions were asked of all household members 15 and older, with some additional qualifying restrictions as noted. Coverage of children under 15 was ascertained from those questions that asked explicitly about the coverage of respondents’ children or other household members--that is, question 23d for Medicaid and questions 24i, 24k and 24l for other types of coverage. Note that question 24i could be the source of reported coverage for adults as well as children.

From the fields provided on the SIPP files we assigned each sample person to one of 13 categories of insurance coverage in each calendar month from January 1992 through September 1994. In making assignments to children, and presumably many spouses as well, we often had to refer to the record of a parent or other adult in the household to determine the source of coverage. This was not true for Medicaid or Medicare, but it was true for every other source. The SIPP data file identifies the household member providing the coverage when the plan is in another member’s name, and we used this information to link to that person’s record and access variables describing the source of coverage. When the coverage was provided by someone outside of the household, no source could be identified. We relegated to a separate category those children whose only coverage was provided by someone outside the household. Similarly, if a child’s coverage was reported under question 24k or 24l, then no information on the source was available.(4) We assigned such children to a residual coverage category indicating that we had no information on the type of coverage.(5)

Finally, a person could have been reported as covered by more than one type of plan during a given month. Clearly, Medicaid could be reported in combination with any other type of plan. Married persons could be reported as having their own employer-sponsored or other plan and being covered by their spouse’s plan. Similarly, children could be reported as covered by both parents’ plans, although the Census Bureau’s coding of the responses does not seem to allow for recording coverage under more than one parent’s plan (there is only one variable pointing to another household member as the source of coverage). Where two or more sources of coverage are reported, they may have been overlapping, or one source may have terminated while another began. Rather than trying to capture and display multiple sources of coverage, we elected to assign a single source of coverage to each person, following a priority scheme. In view of the focus of our research, we assigned Medicaid the highest priority. That is, any child who was reported to be covered by Medicaid, regardless of whatever other coverage may have been reported in that month, was coded as a Medicaid enrollee.(6) Priority was accorded to other coverages in the following order:

  • Medicare
  • Other coverage held in one’s one name
  • Other coverage held in another household member’s name
  • Coverage provided by someone outside the household
  • Generic other coverage not associated with another household member
  • Uninsured

Beyond Medicaid, then, this scheme gives priority to assigning the type of coverage for which we have the most information.

C. Estimates of Health Insurance Coverage

Table 2 reports estimates of the health insurance coverage of children under 19 in each of the first three waves of the 1992 SIPP panel. These results are based on the full panel sample--that is, sample members who were present for the duration of the panel or until such time as they left the SIPP universe. Results are presented for the one month that was common to the reference periods of all four rotation groups in each interview wave--that is, January 1992 for wave 1, May 1992 for wave 2, and September 1992 for wave 3. Presenting Table 2 allows us to introduce the 13 categories of health insurance coverage that we identified with the SIPP variables and to show what change in the distribution of coverage may have occurred between the beginning of the 1992 and the eve of FY93, which along with FY94 is the focal period for the rest of this report.

The upper panel of Table 2 presents estimates of the number of children in each of the 13 coverage categories while the lower panel gives the percentage distribution for all children in each year. Given that the number of children represented by the panel declines over time while the total population of children rises, the percentage distributions provide greater comparability over time and may apply nearly as well to the full population as to the population that the panel actually represents in each month.

Source of Coverage   Wave 1
Jan-92
Wave 2
May-92
Wave 3
Sep-92
TABLE 2: ESTIMATES OF HEALTH INSURANCE COVERAGE OF CHILDREN UNDER 19: FIRST THREE WAVES OF 1992 SIPP PANEL
  TOTAL 72,103,000 71,566,000 71,432,000
1 Medicaid 11,618,000 12,231,000 12,390,000
2 Medicare 0 0 0
Current Employer        
3 Paying All or Part of Cost 39,749,000 39,150,000 39,252,000
4 Paying None of Cost 1,435,000 1,140,000 1,261,000
Former Employer        
5 Paying All or Part 838,000 701,000 744,000
6 Paying None of Cost 329,000 255,000 242,000
7 CHAMPUS 546,000 491,000 452,000
8 CHAMPVA 40,000 33,000 89,000
9 Military 1,140,000 1,091,000 946,000
10 Other 3,120,000 3,128,000 2,938,000
11 Provided by Someone outside the Household 3,197,000 3,044,000 2,837,000
12 Details Unknown 181,000 695,000 483,000
13 Uninsured 9,910,000 9,605,000 9,800,000
  TOTAL 100 100 100
1 Medicaid 16.1 17.1 17.3
2 Medicare 0 0 0
Current Employer        
3 Paying All or Part of Cost 55.1 54.7 55
4 Paying None of Cost 2 1.6 1.8
Former Employer        
5 Paying All or Part 1.2 1 1
6 Paying None of Cost 0.5 0.4 0.3
7 CHAMPUS 0.8 0.7 0.6
8 CHAMPVA 0.1 0 0.1
9 Military 1.6 1.5 1.3
10 Other 4.3 4.4 4.1
         
11 Provided by someone outside the Household 4.4 4.3 4
12 Details Unknown 0.3 1 0.7
13 Uninsured 13.7 13.4 13.7
SOURCE: Survey of Income and Program Participation, 1992 Panel.

The first two sources of coverage, Medicaid and Medicare, require no explanation. Categories 3 through 6 refer to the current or former employer of a child’s parent or guardian, generally, although an older teen could have reported coverage by his or her own employer. Category 3, which accounts for 55 percent of children in each wave, refers to an employer- or union-sponsored plan with the employer or union paying all or part of the cost of the premiums. Category 4, which accounts for 2 percent or less of children, includes plans that the employer or union offers with no subsidization. Category 5 involves coverage that a former employer continues to subsidize while category 6, representing no more than half a percent of all children, is coverage associated with a former employer but with no (further) employer subsidy. Coverage obtained under COBRA would appear to fall in category 6, although the questions on which this category is based could easily lead COBRA participants to report their coverage elsewhere. Categories 7 through 9 refer to coverage provided by CHAMPUS, CHAMPVA, or the military. Category 10 refers to a source of coverage other than 1 through 9. This may include a state-only plan or coverage purchased in the private insurance market. Category 11 is coverage provided by someone living outside the child’s household--typically a divorced or absent parent. SIPP tells us nothing about the source of the coverage because the person in whose name the coverage is held is not interviewed, but we can infer that this coverage would be assigned to one of categories 3 through 10.

Category 12 consists of coverage that could not be classified under one of the preceding 11 categories. This category accounts for only .3 percent of all children in the first wave but then grows to 1 percent in the second wave and remains close to that level in wave 3. We suspected that most of the children classified under category 12 were placed in that category by their parents’ responses to question 24k. That is the children were reported as covered but were not identified with the plan of any adult in the household. If all of the children assigned to category 12 were allotted to the category for that reason, then we would expect to find that all of them were under 15 years of age, based on the final screen for question 24k. While disproportionate numbers of the children in category 12 were indeed under age 15, there remained enough who were over that limit to persuade us that some other explanation was operative. Because the health insurance variables on the SIPP file are not identified with specific questions, it is difficult to determine how a child could end up in category 12 other than by question 24k. We are left to infer that some of the information on children’s coverage must have been incomplete. For example, the Census Bureau may have coded the child as covered under another household member’s plan but either failed to identify or misidentified the household member responsible for the coverage. With further review of individual records it may be possible to resolve why these code 12 assignments were made, but it may not be possible to determine the correct coverage in each instance. Given the low frequency of this category, further investigation may have little merit.

The final category consists of children with no reported coverage. The relative frequency of this category remained at 13.7 percent between waves 1 and 3. During this same period, the proportion of children reporting Medicaid coverage rose from 16.1 percent to 17.3 percent. This rise was offset by a decline in other types of coverage sufficient to leave unchanged the proportion of children reported as uninsured.

Table 3 presents estimates of children’s health insurance coverage for selected months of FY93 and FY94. The first page provides estimates of numbers of children while the second page gives percentages of all children. Reported Medicaid coverage continues the rise observed in Table 2., growing to 19.0 percent of all children by September 1994. The percentage of children who are reported to be without health insurance declines by a full percentage point from September 1992 (in Table 2). The only other notable changes are a decline in the proportion reporting coverage under an employer plan to which the employer makes no contribution and a rise in the proportion of children who are reported to be insured but with no details provided. The former drops from 1.7 percent of all children to 1.2 percent while the latter increases from .8 percent to 1.1 percent.

TABLE 3: HEALTH INSURANCE COVERAGE OF CHILDREN UNDER 19: FY93 AND FY94, SELECTED MONTHS

Source of Coverage Number of Children
Oct-92 Dec-92 Jun-93 Sep-93 Dec-93 Jun-94 Sep-94
    TOTAL 71,549,000 71,323,000 71,111,000 70,868,000 70,795,000 70,197,000 69,935,000
Medicaid 12,614,000 12,829,000 13,328,000 13,369,000 13,233,000 13,264,000 13,253,000
Medicare 0 0 0 0 0 6,000 6,000
Current Employer              
    Paying All or Part of 39,174,000 39,174,000 38,872,000 38,350,000 38,925,000 37,929,000 38,039,000
    Paying None of Cost 1,228,000 1,074,000 1,107,000 968,000 844,000 991,000 837,000
Former Employer              

    Paying All or Part

781,000 639,000 727,000 685,000 790,000 1,031,000 506,000

    Paying None of Cost

276,000 253,000 214,000 246,000 275,000 272,000 309,000
CHAMPUS 539,000 705,000 631,000 567,000 502,000 573,000 465,000
CHAMPVA 89,000 51,000 63,000 79,000 55,000 39,000 88,000
Military 954,000 867,000 864,000 891,000 876,000 829,000 793,000
Other 3,007,000 3,050,000 2,823,000 2,831,000 2,727,000 2,911,000 2,907,000
Provided by Someone outside the Household 2,841,000 2,852,000 2,948,000 3,072,000 2,845,000 2,962,000 3,035,000
Details Unknown 559,000 522,000 587,000 539,000 702,000 670,000 786,000
Uninsured 9,489,000 9,141,000 8,947,000 9,271,000 9,021,000 8,719,000 8,911,000
Source of Coverage Percent of Total Children
Oct-92 Dec-92 Jun-93 Sep-93 Dec-93 Jun-94 Sep-94
    TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Medicaid 17.6 18.0 18.7 18.9 18.7 18.9 19.0
Medicare 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Current Employer              
    Paying All or Part of 54.8 55.2 54.7 54.1 55.0 54.0 54.4
    Paying None of Cost 1.7 1.5 1.6 1.4 1.2 1.4 1.2
Former Employer              
    Paying All or Part 1.1 0.9 1.0 1.0 1.1 1.5 0.7
    Paying None of Cost 0.4 0.4 0.3 0.3 0.4 0.4 0.4
CHAMPUS 0.8 1.0 0.9 0.8 0.7 0.8 0.7
CHAMPVA 0.1 0.1 0.1 0.1 0.1 0.1 0.1
Military 1.3 1.2 1.2 1.3 1.2 1.2 1.1
Other 4.2 4.3 4.0 4.0 3.9 4.1 4.2
Provided by Someone outside the Household 4.0 4.0 4.1 4.3 4.0 4.2 4.3
Details Unknown 0.8 0.7 0.8 0.8 1.0 1.0 1.1
Uninsured 13.3 12.8 12.6 13.1 12.7 12.4 12.7
SOURCE: Survey of Income and Program Participation, 1992 Panel.

TABLE 4: CHILDREN UNDER 19 ENROLLED IN MEDICAID OR WITHOUT HEALTH INSURANCE AT END OF FISCAL YEAR OR EVER IN YEAR

Health Insurance by Period Number of Children Under 19 by Health Insurance Total Children Under 19 in Period Percent of Total Children Under 19
Medicaid      
    Enrolled in September 1993 13,369,000 70,868,000 18.9%
    Ever enrolled in FY93 17,800,000 74,691,000 23.8%
    Enrolled in September 1994 13,259,000 69,935,000 19.0%
    Ever enrolled in FY94 17,795,000 73,619,000 24.2%
    Ever enrolled in FY93 or FY94 21,043,000 77,697,000 27.1%
Uninsured      
    Uninsured in September 1993 9,271,000 70,868,000 13.1%
    Ever uninsured in FY93 16,089,000 74,691,000 21.5%
    Uninsured in September 1994 8,911,000 69,935,000 12.7%
    Ever uninsured in FY94 15,936,000 73,619,000 21.6%
    Ever uninsured in FY93 or FY94 21,074,000 77,697,000 27.1%
SOURCE: Survey of Income and Program Participation, 1992 Panel.

The upper panel of Table 4 compares enrollment in Medicaid at the end of each fiscal year to enrollment ever during the year. The lower panel makes the same comparison for uninsurance. For Medicaid the number ever enrolled in each fiscal year is one-third higher than the number enrolled at the end of the year. The number of children ever enrolled in Medicaid during the full two-year period is 58 percent higher than the number enrolled at the end of either year.(7) When compared to the one-year ever-enrollment, the two-year figure indicates very modest turnover from one year to the next, as the number of children ever enrolled over two years is less than one-fifth higher than the number ever enrolled in the first or second year.

Turnover among the uninsured is considerably higher than it is among Medicaid enrollees. The number of children ever uninsured during a fiscal year is about 75 percent higher than the number who are uninsured at the end of the year. The number ever uninsured during a two-year period is about 30 percent higher than the number ever uninsured in either year alone and about 130 percent higher than the number uninsured at the end of either year (or at any one time). For the two-year period, the percentage of children ever uninsured matches the fraction who were ever on Medicaid: 27.1 percent.

Table 5 presents estimates of the number and percentage of children who were without health insurance coverage by month for the calendar years 1992 through 1994. The percentage uninsured declines over the course of the first year but shows no trend after that. There is no evidence of seasonality in these numbers. We wanted to select two or three months to present statistics in this and the other technical appendices, and this table suggests that the choice is not particularly important. We elected to focus on the final two months of FY93 and FY94 in most of our cross- sectional tables, supplementing these with the first month of FY93 when appropriate. September 1993 happens to have a relatively high estimate of the uninsured, at 13.1 percent, compared to other months whereas September 1994, at 12.7 percent, is close to the average. It is important to keep in mind that the month-by-month results suggest that the difference between these two estimates does not reflect any trend.

TABLE 5: NUMBER AND PERCENTAGE OF CHILDREN UNDER 19 WHO WERE WITHOUT HEALTH INSURANCE COVERAGE, BY MONTH: 1992 THROUGH 1994

  Thousands of Children Percent of All Children
Month 1992 1993 1994 1992 1993 1994
January 9,910 9,126 9,009 13.7 12.8 12.7
February 9,950 9,144 9,098 13.8 12.9 12.9
March 9,790 8,974 9,050 13.6 12.6 12.8
April 9,493 9,133 8,998 13.3 12.8 12.8
May 9,605 9,137 9,025 13.4 12.8 12.8
June 9,610 8,947 8,719 13.4 12.6 12.4
July 9,639 9,222 8,694 13.5 13.0 12.4
August 9,713 9,134 8,745 13.6 12.9 12.5
September 9,800 9,271 8,911 13.7 13.1 12.7
October 9,489 9,080 9,017 13.3 12.8 12.9
November 9,238 8,920 8,966 12.9 12.6 12.9
December 9,141 9,021 8,749 12.8 12.7 12.6
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Table 6 presents similar figures for reported Medicaid coverage. Here it is notable that the reported coverage rises from 16.1 percent in January 1992 to 18.7 percent in June 1993, then levels off (with modest swings). It is notable that the percentage point rise in reported Medicaid enrollment is nearly three times the decline in the percentage of children reported to be without health insurance coverage. This suggests that two-thirds of the enrollment increase is due to net movement from other sources of coverage rather than net movement out of the uninsured. We address this issue more directly in Technical Appendix B.

TABLE 6: NUMBER AND PERCENTAGE OF CHILDREN UNDER 19 WHO WERE REPORTED TO BE COVERED BY MEDICAID, BY MONTH: 1992 THROUGH 1994
  Thousands of Children Percent of All Children
Month 1992 1993 1994 1992 1993 1994
January 11,618 12,842 13,231 16.1 18 18.7
February 11,711 12,901 13,228 16.3 18.1 18.7
March 11,981 13,187 13,079 16.7 18.5 18.5
April 12,319 13,264 13,185 17.2 18.6 18.7
May 12,231 12,983 13,109 17.1 18.2 18.6
June 12,373 13,328 13,270 17.3 18.7 18.9
July 12,593 13,294 13,372 17.6 18.7 19.1
August 12,518 13,233 13,418 17.5 18.6 19.2
September 12,390 13,369 13,259 17.3 18.9 19.0
October 12,614 13,236 12,946 17.6 18.7 18.5
November 12,800 13,263 12,830 17.9 18.7 18.4
December 12,829 13,233 12,815 18.0 18.7 18.4
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Table 7 provides information on sources of health insurance coverage in FY93, FY94, and the combined, two-year period. Figures for FY93 refer to coverage at any time during that year, and likewise for the FY94 figures. Figures for “FY93 and FY94" refer to the two year period. These estimates are based on our monthly measure of insurance coverage. While a child may in fact have been covered by more than one source in a month (or even at the same time), and SIPP can tell us about multiple sources of coverage during a month, we measured only one source per month, as explained above. Persons with multiple sources in this table, therefore, were covered by those sources at different times during the year (or two-year period for the last two columns of the table).

TABLE 7: PATTERNS OF HEALTH INSURANCE COVERAGE AMONG CHILDREN UNDER 19: FY93 AND FY94
  FY93 FY94 FY93 and FY94
Coverage Number Percent Number Percent Number Percent
    All Children under 19 74,691,000 100 73,619,000 100 77,697,000 100
Children with No Coverage during the Period 5,001,000 6.7 4,731,000 6.4 3,569,000 4.6
Children with Any Coverage during the Period 69,690,000 93.3 68,888,000 93.6 74,128,000 95.4
    Children with only one source of coverage 62,995,000 84.3 62,442,000 84.8 62,554,000 80.5
        Employer-sponsored 47,187,000 63.2 46,366,000 63 46,921,000 60.4
        Medicaid 13,918,000 18.6 14,124,000 19.2 14,075,000 18.1
        Other 1,889,000 2.5 1,952,000 2.7 1,558,000 2
    Children with multiple sources of coverage 6,695,000 9 6,446,000 8.8 11,574,000 14.9
        Employer-sponsored and Medicaid 3,550,000 4.8 3,373,000 4.6 6,242,000 8
        Employer-sponsored and other 2,813,000 3.8 2,775,000 3.8 4,606,000 5.9
        Medicaid and other 225,000 0.3 171,000 0.2 349,000 0.4
        All three sources 107,000 0.1 127,000 0.2 376,000 0.5

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: All children under 19 includes children who were under 19 at the start of the indicated period as well as children born during the period.

We note that 6.7 percent of children reported no coverage of any kind in FY93, 6.4 percent reported no coverage in FY94, and 4.6 percent reported no coverage in either year. The remainder of the children in each of these periods reported coverage for at least part of the period. Between 84 and 85 percent reported only one source of coverage in each of the two years while 80.5 percent reported only one source over the two-year period. Within each of the two years, 63 percent reported only employer-provided coverage while about 19 percent reported only Medicaid coverage, with about 2.5 percent reporting coverage from another source. For the two-year period, these figures changed little: 60 percent reported only employer-provided insurance, 18 percent reported only Medicaid, and 2 percent reported only another source of coverage. We know from Table 3 that the proportion of children ever covered by Medicaid in the two-year period was higher than the proportion covered in either year alone. Table 7 shows that some of this increase is due to more children reporting Medicaid along with another source over the two-year period relative to one year. Specifically, 9 percent of children report Medicaid in combination with one or two other sources over the two-year period compared to 5 percent for either year alone. (The rest of the increase in Medicaid enrollment must be among children who were otherwise uninsured.)

In Table 8 we look at the incidence of uninsurance among children who reported any coverage during either or both fiscal years. Among children with any coverage, about 16 percent reported one or months of uninsurance in either fiscal year, and 24 percent reported a spell of uninsurance over the two-year period. The probability that a child with any coverage during a period was ever uninsured during that period varies substantially by type of coverage. Among children who were ever covered by Medicaid in either year, 28 percent were uninsured for part of the year. Among those who were ever covered by Medicaid over the two year period, the percentage who were ever uninsured during that period was 41 percent. For those with employer-provided insurance, the percentages uninsured were less than half these figures. Among those with other coverage, the proportions with any months of uninsurance were about midway between those for children ever covered by Medicaid versus employer-provided insurance.

Source of Coverage FY93 FY93 and FY94
TABLE 8: FREQUENCY OF UNINSURANCE AMONG CHILDREN WITH ANY COVERAGE, BY SOURCE: FY93 AND FY94
Children with any coverage 69,690,000 68,888,000 74,128,000
    Number ever uninsured 11,088,000 11,205,000 17,504,000
    Percent ever uninsured 15.9% 16.3% 23.6%
    Medicaid 17,800,000 17,795,000 21,043,000
        Number ever uninsured 5,077,000 5,002,000 8,569,000
        Percent ever uninsured 28.5% 28.1% 40.7%
    Employer-provided 53,657,000 52,641,000 58,146,000
        Number ever uninsured 6,750,000 6,852,000 11,627,000
        Percent ever uninsured 12.6% 13.0% 20.0%
    Other coverage 5,034,000 5,025,000 6,890,000
        Number ever uninsured 975,000 1,032,000 2,031,000
        Percent ever uninsured 19.4% 20.5% 29.5%
Children with one source of coverage 62,995,000 62,442,000 62,554,000
    Number ever uninsured 9,387,000 9,543,000 12,934,000
    Percent ever uninsured 14.9% 15.3% 20.7%
    Medicaid 13,918,000 14,124,000 14,075,000
        Number ever uninsured 3,808,000 3,825,000 5,163,000
        Percent ever uninsured 27.4% 27.1% 36.7%
    Employer-provided 47,187,000 46,366,000 46,921,000
        Number ever uninsured 5,150,000 5,300,000 7,276,000
        Percent ever uninsured 10.9% 11.4% 15.5%
    Other coverage 1,889,000 1,952,000 1,558,000
        Number ever uninsured 429,000 418,000 495,000
        Percent ever uninsured 22.7% 21.4% 31.8%
SOURCE: Survey of Income and Program Participation, 1992 Panel.

If we look at just those children who report one source of coverage, so that we can isolate the “impact” of the source of coverage, we find even more pronounced differences by type of coverage. While the incidence of uninsurance is generally lower among children with only one source versus two or more, children whose only source was Medicaid during a year had a 27 percent incidence of uninsurance while those whose only source was employer-provided insurance had only an 11 percent incidence of uninsurance. Among those with coverage from another source during a year, between 21 and 23 percent were uninsured for at least one month in the year.

These figures are particularly interesting in light of the policy focus on the CPS, which counts as uninsured only those persons who reported having had no coverage during the year and provides no estimate of persons who were uninsured for only part of the year. Estimates of the kind reported in Table 8 cannot be constructed with CPS data

Technical Appendix B: The Dynamics of Uninsurance, Medicaid Eligibility and Participation among Children: 1993-1994

This report uses data from the Survey of Income Program Participation (SIPP) to investigate dynamic aspects of the lack of health insurance, eligibility for Medicaid, and Medicaid participation among children in 1993 and 1994. After describing our data source we present findings on the duration of spells of uninsurance based on an analysis of new spells as well as spells active at a point in time. We explain why these alternative approaches to measuring the duration of uninsurance yield such different distributions of spell length. We also present estimates of the amount of time that uninsured children are eligible for Medicaid. From there we turn to an examination of the patterns of health insurance coverage preceding and following completed spells of uninsurance, and we compare spells of different lengths. We do this for spells of Medicaid-eligible uninsurance as well. Next we examine the frequency of transitions in coverage, beginning with changes in coverage between one year and the next and then turning our attention to monthly transitions in uninsurance and Medicaid enrollment, culminating in the presentation of estimates of cumulative monthly exit rates from uninsurance, Medicaid-eligible uninsurance, non-Medicaid-eligible uninsurance, and Medicaid enrollment. In the closing sections we discuss churning versus turnover as phenomena in the dynamics of uninsurance and Medicaid enrollment and then summarize our principal findings.

A. Data

The findings presented herein are based on an analysis of longitudinal data from the 1992 panel of the SIPP. Starting with an initial sample of about 20,000 households representing the resident, noninstitutionalized population of the United States, except for those living in military barracks, the Census Bureau reinterviewed respondents every four months over a period of three calendar years. Respondents were asked a lengthy set of questions about their labor force participation, sources and amounts of income, family and household composition, and participation in government programs in each of the preceding four months. The interviews were staggered, with one quarter of the sample, constituting a “rotation group,” being interviewed each month. The purpose of this design was, in part, to minimize the impact of recall error and other reporting problems on the quality of the data compiled for calendar months. The estimates for a given calendar month are based on roughly equal proportions of the respondents describing circumstances one month ago, two months ago, three months ago, and four months ago.

Several features of the SIPP make these data especially appealing for the analysis of dynamic behavior in children’s health insurance coverage. The SIPP provides a detailed measure of health insurance coverage for every month of the two to three year duration of a panel. Measures of the duration of new spells of particular types of coverage or lack of coverage can be constructed by aggregating the reports from successive interviews, so that no matter how long the measured duration of a particular spell, no part of a reported spell relies on respondent recall beyond four months. In addition to providing measures of health insurance coverage, SIPP also provides very detailed measures of demographic and economic characteristics--again on a monthly basis. This affords us the opportunity to construct contemporaneous measures of circumstances that may affect eligibility for and enrollment in particular types of coverage. For example, we can construct a reasonably good simulation of Medicaid eligibility on a monthly basis and compare these monthly estimates to reports of Medicaid participation and insurance coverage--or the lack thereof--on a monthly basis as well. Arguably, there are no better data for investigating the dynamics of children’s health insurance coverage in the context of changing circumstances, including Medicaid eligibility.

Despite these strengths, however, the SIPP does possess some notable limitations that will become evident as we proceed through our findings. These include some well-documented but not well understood problems with the quality of retrospective monthly reports of program participation and health insurance coverage.

B. Duration of Uninsurance

The length of time that children spend without health insurance coverage is exceedingly important to policymakers. Brief spells of uninsurance carry very different policy implications than long spells. While transitional spells of even one month in duration carry risks to the temporarily uninsured, the policy responses that they invite are clearly different than those demanded to address chronic lack of insurance. The advent of the Children’s Health Insurance Program (CHIP) gives heightened importance to the duration of uninsurance as a policy variable. Federal law allows states considerable freedom in determining how much of the uninsured population to cover and how to define eligibility. A number of states are specifying minimum periods of uninsurance before a child can qualify for the new coverage.

How long do spells of uninsurance last? Here we present findings with respect to the duration of spells without insurance. We contrast two different ways of identifying spells: new spells that began within a particular 12-month period and active spells that were in progress at the end of that period. In addition to looking at spells of uninsurance, we examine spells in which the uninsured were eligible for Medicaid and, theoretically, could have availed themselves of such coverage.

1. New Spells

Table 1 describes the duration of spells of uninsurance that began during FY93. For all of these spells, the first set of columns in Table 1 gives the distribution of completed duration in single months, up to a maximum of 13 or more. For most of this last category we did not observe the spell completion; we know only that these spells lasted at least 13 months. We combine all spells of 13 months or greater because 13 is the longest duration that we can observe with the SIPP panel for a spell starting in the final month of FY93. To produce representative estimates of longer durations, we must limit our sample to spells starting earlier in the year. For spells starting in the first six months of FY93, therefore, the second set of columns in Table 1 reports completed durations up to 19 months or more.1

TABLE 1: SPELLS OF UNINSURANCE STARTING IN FY93 BY COMPLETED DURATION IN MONTHS
  Spells Starting 10/92 to 9/93 Spells Starting 10/92 to 3/93
Completed Duration of Spell Number Percent Cumulative Percent Number Percent Cumulative Percent
All Spells 9,205,000 100   4,094,000 100  
1 Month 683,000 7.4 7.4 204,000 5 5
2 Months 638,000 6.9 14.3 302,000 7.4 12.4
3 Months 485,000 5.3 19.6 211,000 5.2 17.5
4 Months 3,166,000 34.4 54 1,375,000 33.6 51.1
5 Months 299,000 3.2 57.3 217,000 5.3 56.4
6 Months 223,000 2.4 59.7 123,000 3 59.4
7 Months 237,000 2.6 62.3 66,000 1.6 61
8 Months 798,000 8.7 70.9 449,000 11 72
9 Months 57,000 0.6 71.5 25,000 0.6 72.6
10 Months 78,000 0.9 72.4 22,000 0.5 73.2
11 Months 108,000 1.2 73.6 56,000 1.4 74.5
12 Months 584,000 6.3 79.9 210,000 5.1 79.7
13+ Months 1,848,000 20.1 100 833,000 20.3 100
13 Months       6,000 0.1 79.8
14 Months       22,000 0.5 80.3
15 Months       28,000 0.7 81
16 Months       128,000 3.1 84.2
17 Months       5,000 0.1 84.3
18 Months       8,000 0.2 84.5
19+ Months       636,000 15.5 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Perhaps the most striking feature of these distributions is the heaping of reported durations at four months. Of the 9.2 million spells that began in FY93, more than one-third were completed in exactly four months. This is not a true reflection of spell durations but a phenomenon of the measurement process in the SIPP. This heaping of responses at four months, and to a lesser degree at multiples of four months (there are local peaks at 8, 12, and 16 months), is the result of a “seam effect.” As we mentioned above, SIPP respondents are interviewed every four months and asked extensive questions about their circumstances in each of the four preceding months. For reasons that are not fully understood, transitions in many characteristics are reported as occurring dispro- portionately between the four-month reference periods of interviews--that is, at the seams--rather than within these reference periods. It happens that this seam effect is particularly strong for transitions into and out of insurance coverage. For example, while we would expect only one quarter of new spells of uninsurance to begin in the first month of a reference period or end in the last month, 75 to 90 percent of new spells are reported as beginning in the first month or ending in the fourth month. The transitions, in other words, are reported to have occurred between the reference periods of the interviews. Explanations for this phenomenon include genuine recall error as well as various kinds of behaviors that respondents or interviewers may engage in to speed up or simplify the interview. Clearly, treating the previous four months as a single reporting period--for at least some kinds of responses--simplifies the interview, but if that is what is happening in the reporting or recording of health insurance coverage, then we have to interpret the data accordingly.

Because of the staggered interviewing, described earlier, the seam effect does not show up in the estimates reported for calendar months. But this is not to say that the seam effect does not influence the overall magnitude of these calendar month or point-in-time estimates. Unless respondents (or interviewers) are as likely to “round up” or “round down” their reported months of participation in a reference period, the monthly estimates of aggregate participation will be affected. In a later report, we consider how the seam effect might affect estimates of the number of children who are uninsured.

The seam effect is not unique to SIPP, by any means, and the research that preceded the initiation of the first SIPP panel in 1984 included an examination of this phenomenon. Panel surveys with annual interviews find the same problem when they attempt to measure monthly behavior. To the extent that recall error contributes to the seam effect in SIPP, the impact of such error will be magnified in surveys that attempt to measure behavior over longer periods of time--such as an entire year, which is the reference period for much of the data collected in the March supplement to the Current Population Survey (CPS). Indeed, the fact that CPS estimates of the number of people who had no health insurance coverage during the previous calendar year resemble point-in-time estimates more closely than they resemble the calendar year estimates produced by panel surveys is indicative of a serious response problem that may share some underlying causes with the SIPP seam effect.

Nevertheless, it is clear that we have to take account of the seam effect when we interpret data on spell durations, such as those reported in Table 1. Estimates of the duration of completed spells reflect the maximum impact of the seam effect because each reported spell is subject to this influence at both ends. Of the nearly 3.2 million spells of uninsurance reported as being completed in exactly four months, how many were really completed in four months, and how many were completed in three or five months, two or six months, or even one month or seven months? And how many spells of one, two, or even three months duration were not reported at all?

While there may be ways to tease out more information from the SIPP data, we have not attempted to do so here.2 Instead we simply recognize the limitations of these data. Specifically, despite what the SIPP data appear to tell us about the length of short spells of uninsurance, we would suggest that these data should not be used to draw inferences about the relative frequency of spell lengths shorter than six months or about longer spell lengths in groupings of fewer than four months. In other words, it may be reasonable to infer from Table 1 that about 60 percent of new spells of uninsurance in FY93 were completed in six months or less, but the data will not support inferences about how many of these spells may have been completed in two months or less, or four months or less. Similarly, it may be correct to infer that about 8 percent of spells lasted 9 to 12 months or that 3.5 to 4 percent lasted 15 to 18 months but certainly not that only .1 percent were completed in exactly 13 months. In general, we need to be conscious that the numbers reported at 4 months, 8 months, 12 months, or 16 months almost certainly include more spells of shorter or longer lengths than they do spells of the nominal length.

Spell lengths of six months or longer or 12 months or longer are becoming important in defining eligibility for CHIP coverage in a number of states. What can we infer about the relative frequency of such spells from the data presented in Table 1? Read literally, the data in Table 1 tell us that about 43 percent of the spells that began in FY93 lasted six months or longer (100 minus the 57.3 percent with lengths of five months or less), and 26.4 percent of the new spells in that year lasted 12 months or longer. Recognizing that most of the spells reported as having durations of exactly four or 12 months were actually longer or shorter than that, we would suggest that the true proportion of new spells with durations of six months or longer could be as much as 5 percentage points higher--up to 48 percent--and that the true proportion of spells with durations of 12 months or longer could be 2 to 3 percentage points lower than the observed figure, or closer to 24 percent.

2. Active Spells

Where Table 1 presented distributions of new spells of uninsurance that began in FY93, Table 2 presents distributions of all spells that were active at the end of FY93. For all children who were uninsured in September 1993, Table 2 tells us how long they had been uninsured. The distribution of these “current” durations can be compared to the distribution of completed durations of new spells, which is reproduced from Table 1 in the right hand columns of Table 2.

It is important to understand how the durations of active spells reported in the left side of Table 2 differ from the durations of new spells reported in the right side (and in Table 1). First, the durations of active spells are incomplete. More specifically, they are “right censored.” If we think of a duration as spanning from left to right on a time line, in other words, what we do not observe for these active spells is how far their durations extend to the right. All of the spells terminate at the same point in time--September 1993. We do not know how long these durations will be when they are completed. The second respect in which the active spells portrayed in Table 2 differ from the completed spells presented earlier is that they share a common month. All of the active spells include September 1993. While there is a lot of overlap in the spells reported in Table 1, these spells do not share a common month.

TABLE 2: SPELLS OF UNINSURANCE ACTIVE IN SEPTEMBER 1993 AND SPELLS STARTING IN FY93 BY CURRENT OR COMPLETED DURATION IN MONTHS
  Spells Active in September 1993 Spells Starting 10/92 to 9/93
Duration of Spell Number Percent Cumulative Percent Number Percent Cumulative Percent
All Spells 9,271,000 100   9,205,000 100  
1 Month 1,006,000 10.9 10.9 683,000 7.4 7.4
2 Months 799,000 8.6 19.5 638,000 6.9 14.3
3 Months 842,000 9.1 28.6 485,000 5.3 19.6
4 Months 561,000 6.1 34.6 3,166,000 34.4 54
5 Months 259,000 2.8 37.4 299,000 3.2 57.3
6 Months 339,000 3.7 41.1 223,000 2.4 59.7
7 Months 229,000 2.5 43.5 237,000 2.6 62.3
8 Months 292,000 3.1 46.7 798,000 8.7 70.9
9 Months 267,000 2.9 49.6 57,000 0.6 71.5
10 Months 169,000 1.8 51.4 78,000 0.9 72.4
11 Months 179,000 1.9 53.3 108,000 1.2 73.6
12 Months 81,000 0.9 54.2 584,000 6.3 79.9
13+ Months 4,247,000 45.8 100 1,848,000 20.1 100
13 Months 202,000 2.2 56.4      
14 Months 102,000 1.1 57.5      
15 Months 148,000 1.6 59.1      
16 Months 116,000 1.2 60.3      
17 Months 129,000 1.4 61.7      
18 Months 120,000 1.3 63      
19+ Months 3,432,000 37 100      

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: Reported durations for active spells represent current durations as of September 1993. Reported durations for spells starting during FY93 represent completed durations.

These two differences have important implications for how the distributions of durations differ. First, let us consider what the durations reported on the left side of Table 2 represent. Durations of one month represent spells that started in September. Ultimately, very few of these spells will end at one month. Instead, if we could follow them through to completion, we would find that they have essentially the same distribution as the spells reported in Table 1. Durations of two months refer to all spells that started in August 1993, minus those that ended in that same month--that is, spells with completed durations of one month. If we were to follow these spells to completion, they too would have the same distribution as the spells reported in Table 1 except for the complete absence of spells of one month duration. Similarly, the 12-month durations include all spells that started in October 1992 minus those that ended in any of the months from October 1992 through August 1993.3

What we notice first in the distribution of active spells is the absence of a seam effect. There are no peaks whatsoever at multiples of four months. This absence of a seam effect is due entirely to the right censoring. Most of these spells are not complete, so the seam effect is hidden. One quarter of the spells that will eventually end in four months shows up as one month in length, one quarter shows up as two months in length, one quarter appears as three, and one quarter as four months in length. Far from producing peaks at four months, the seam effect actually contributes to the relative evenness of the distribution of spells across durations of one to four months.

The second thing we notice is that apart from the higher proportion of spells with durations of three months or less, the spells that are active at the end of the year (and this would be true of spells that were active during any given month of the year) are longer, on average, than the spells that started during the year--even though most of the active spells are not yet complete. The median length of the active spells lies between 9 and 10 months whereas the median completed length of the new spells lies close to 4 months. Nearly 46 percent of the active spells have durations in excess of 12 months compared to just 20 percent of the spells that started during the year.

If we followed the active spells to their completion, we would obtain a distribution that is heavily skewed toward long durations. Why is this? It turns out that the representation of spells among active spells is directly proportional to their completed duration. Spells of one month duration are represented solely by spells that began in a single month: September 1993. Spells of two months duration are represented solely by spells that began in either of two months: August or September. Spells of three months duration are represented by spells that began in any of three months, and so on. Thus spells of 12 months duration are represented by spells that began in any of 12 months--October 1992 through September 1993--while spells of 36 months duration are represented by spells that began in any of 36 months--from October 1990 through September 1993. Compared to a distribution of spells starting in the same month or the same year, therefore, spells of exactly 36 months in length are represented at 36 times their relative frequency.

We could estimate the distribution of completed durations among spells that were active in September 1993 by taking the completed distribution of new spells reported in Table 1 and multiplying the number of one-month spells by 1, the number of two-month spells by 2, number of three-month spells by 3, and so on. Obviously, we cannot complete this exercise with the data presented in Table 1 because we do not know the completed duration of spells that exceeded 18 months in length. But if we make an assumption about the average completed duration of spells that were completed in more than 18 months, then we can estimate the relative frequency of such spells among active spells.

Table 3 displays this projected distribution of the completed duration of spells that were active in September 1993. These calculations assume, of course, that the distribution of completed durations did not change over time, so that the relative frequency of, say, an 18-month spell is the same for spells starting at any time between October 1991 and September 1993. Whether or not this assumption is satisfied, the projected distribution illustrates our point. In Table 3 we see that only 22 percent of these spells are completed in six months or less compared to about 60 percent of the new spells in Table 1. At the other end of the distribution, 49 percent of the spells reported in Table 3 extend beyond 18 months compared to only 15.5 percent of the spells in Table 1.

If these different ways of measuring the duration of uninsurance yield such different distributions, can they all be useful? The answer is yes, but the different distributions address different questions. Table 1 reflects a representative sampling of spells. It tells us that most spells of uninsurance among children are relatively short in length, 60 percent being six months or less. The distribution has a long tail, however, with at least 20 percent of spells (we suggested somewhat higher) exceeding 12 months in length and about three-fourths of these surpassing 18 months. It is this long tail that makes possible the skewed distribution that we see when we look at the spells of a representative sample of people with spells active in a give month. People who experience long spells are more likely to be uninsured in a given month than people who experience short spells, so children who are uninsured at a point in time exhibit disproportionately many long spells. Table 3 suggests that about half of the 9 million children who were uninsured in any given month of FY93 will have had spells of more than 18 months in length by the time they become insured (again), and most of these spells will extend well beyond 18 months. Finally, Table 2 addresses a question that is becoming very pertinent with the introduction of CHIP--namely, if eligibility is limited to children who have been uninsured for a specified number of months, how many children will satisfy these criteria? Table 2 indicates that with a 12 month minimum, about 47 percent of all uninsured children would meet this requirement. The percentage among children who meet the income and other requirements is likely to be higher. Lowering the duration requirement to 6 months would add another 16 percent of children--a relatively modest increase given that the minimum duration would be cut in half.

TABLE 3: PROJECTED COMPLETED DURATION OF SPELLS OF UNINSURANCE THAT WERE ACTIVE IN A GIVEN MONTH OF FY93
Projected Completed Duration of Spell Percent Cumulative Percent
All Spells 100  
1 Month 0.8 0.8
2 Months 1.5 2.3
3 Months 1.7 4
4 Months 14.8 18.8
5 Months 1.7 20.5
6 Months 1.6 22.1
7 Months 1.9 24
8 Months 7.5 31.5
9 Months 0.6 32.1
10 Months 0.9 33
11 Months 1.4 34.4
12 Months 8.2 42.6
13+ Months 57.4 100
13 Months 0.2 42.8
14 Months 0.8 43.6
15 Months 1.1 44.7
16 Months 5.3 50
17 Months 0.2 50.2
18 Months 0.4 50.6
19+ Months 49.4 100

SOURCE: Mathematica Policy Research, Inc.

NOTE: In calculating these projections from the distribution of new spells, as reported in Table 1, we assumed that spells of more than 18 months' duration had a mean duration of 30 months. Lowering this assumed mean duration to 24 months would reduce the share of spells with durations of 19+ months to 43.9 percent and increase the cumulative share of spells with durations less than 12 months to 38.2 percent.

3. Medicaid-Eligible Uninsurance

The large number of uninsured children who appear to be eligible for Medicaid has captured the attention of policymakers and advocacy groups alike. In Technical Appendix A we present estimates that 30 to 33 percent of the children who were uninsured at a point in time in FY93 or FY94 were eligible for Medicaid, based on an eligibility simulation that accounted for a substantial majority but certainly not all Medicaid-eligible children. Here we ask how long children who are uninsured and eligible for Medicaid tend to remain both uninsured and Medicaid-eligible.

Table 4 provides information on the duration of spells of Medicaid-eligible uninsurance. A spell of Medicaid-eligible uninsurance begins in the first month that a child is both uninsured and Medicaid-eligible. A child may have been uninsured for one or more months before becoming Medicaid-eligible or may have been Medicaid-eligible at the start of a spell of uninsurance. Likewise, a child may have lost Medicaid eligibility before becoming insured or may have remained eligible through the end of a spell of uninsurance.4 Because patterns may vary, we cannot infer the full length of a spell of uninsurance from the number of months that a child was both Medicaid eligible and uninsured.

TABLE 4: SPELLS OF MEDICAID-ELIGIBLE UNINSURANCE STARTING IN FY93 BY COMPLETED DURATION
  Spells Starting 11/92 to 9/93 Spells Starting 11/92 to 3/93
Completed Duration of Spell Number Percent Cumulative Percent Number Percent Cumulative Percent
All Spells 6,814,000 100   3,161,000 100  
1 Month 1,801,000 26.4 26.4 819,000 25.9 25.9
2 Months 1,140,000 16.7 43.2 496,000 15.7 41.6
3 Months 708,000 10.4 53.5 344,000 10.9 52.5
4 Months 1,450,000 21.3 74.8 634,000 20.1 72.6
5 Months 293,000 4.3 79.1 179,000 5.7 78.2
6 Months 188,000 2.8 81.9 104,000 3.3 81.5
7 Months 79,000 1.2 83 38,000 1.2 82.7
8 Months 295,000 4.3 87.4 155,000 4.9 87.6
9 Months 56,000 0.8 88.2 6,000 0.2 87.8
10 Months 81,000 1.2 89.4 42,000 1.3 89.1
11 Months 46,000 0.7 90.1 34,000 1.1 90.2
12 Months 255,000 3.7 93.8 77,000 2.4 92.6
13+ Months 422,000 6.2 100 233,000 7.4 100
13 Months       11,000 0.4 93
14 Months       26,000 0.8 93.8
15 Months       8,000 0.3 94.1
16 Months       51,000 1.6 95.7
17 Months       0 0 95.7
18 Months       24,000 0.7 96.4
19+ Months       112,000 3.6 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

We began our Medicaid-eligibility simulation in October 1992, so we cannot identify new spells of Medicaid-eligible uninsurance starting before November 1992. (We need to observe the previous month to determine if a new spell has started.) The spells reported in Table 4, therefore, began in November 1992 or any of the subsequent months through September 1993.

Spells of Medicaid-eligible uninsurance tend to be much shorter than spells of uninsurance generally. More than one-half of all spells started in FY93 were completed in three months or less, and only 6 percent extended beyond one year. Among all spells of uninsurance started during the same period, 20 percent reached beyond one year, and more than 15 percent extended at least another six months beyond that.

Finally, the SIPP seam effect is much less pronounced for spells of Medicaid-eligible uninsurance than for all spells of uninsurance. We attribute this to the fact that simulated Medicaid- eligibility is not a characteristic that is reported by survey respondents or recorded by survey interviewers but a complex construct of a great many variables that change at different rates and times. Nevertheless, there are local peaks at four, eight, and 12 months. Taking this into account suggests that the proportion of spells of Medicaid-eligible uninsurance completed in three months or less could be as high as 60 percent.

C. Coverage Before and After Spells of Uninsurance

Other than children who are uninsured at birth, children who begin spells of uninsurance do so by losing previous coverage. What is the source of this coverage? Does it mirror the coverage of all insured children, or do the uninsured come disproportionately from former Medicaid enrollees? Similarly, when children regain coverage do they resemble the rest of the insured population with respect to the sources of this coverage or do they rely on a different mix of sources? And do children tend to return to the same type of coverage that they had prior to their spell of insurance, or do they often end up with another source of coverage? We address these questions by comparing the source of coverage before and after completed spells of uninsurance. We also examine how these answers differ by the duration of spells of uninsurance. After addressing these question with respect to spells of uninsurance, we examine spells of Medicaid-eligible uninsurance to determine what precedes and follows these spells. Here, however, a spell need not be preceded by coverage; instead a child may enter a spell of Medicaid-eligible uninsurance from the state of having been uninsured but not eligible for Medicaid, and may leave a spell of Medicaid-eligible uninsurance by losing Medicaid eligibility without regaining any form of coverage.

1. Spells of Uninsurance

Table 5 presents for spells of uninsurance of various durations a cross-tabulation of insurance coverage in the month before the spell began by insurance coverage in the month after the spell ended (that is, the first month of coverage). Separate tabulations are reported, first, for all spells completed in 1 to 12 months, followed by tabulations of spells completed in 1 to 4 months, 5 to 8 months, and 9 to 12 months. A tabulation of spells completed in 13 to 18 months, derived from spells that started in just the first half of FY93, is reported at the bottom of the table. We did not pool this subset of spells with those completed in 12 months or less because the spells of longer duration would be underrepresented in the pooled tabulation.5

  Insurance Coverage After Completion of Spell
Insurance Coverage Before Start of Spell Employer- sponsored Medicaid Other Not in Universe Total
TABLE 5: INSURANCE COVERAGE BEFORE AND AFTER COMPLETED SPELLS OF UNINSURANCE BEGINNING IN FY93, BY DURATION OF SPELL
  All Spells Completed in 12 Months or Less
Employer-sponsored 41.8 9.4 2.6 0.3 54.1
Medicaid 8.5 28.3 0.9 1.1 38.8
Other 2 1 1.5 0 4.5
Not in Universe 0.6 1.7 0.1 0.1 2.5
Total 52.9 40.5 5.1 1.6 100
  Spells Completed in 1 to 4 Months
Employer-sponsored 43.7 8.9 1.8 0.4 54.7
Medicaid 8.2 28.7 1 1.6 39.5
Other 1.5 0.7 1.5 0 3.6
Not in Universe 0.4 1.5 0.1 0.2 2.2
Total 53.7 39.8 4.4 2.1 100
  Spells Completed in 5 to 8 Months
Employer-sponsored 39.6 10.8 5.1 0.4 55.9
Medicaid 5.3 28.5 0.4 0 34.1
Other 2.9 1.6 2.1 0 6.5
Not in Universe 1.8 1.6 0 0 3.4
Total 49.5 42.5 7.6 0.4 100
  Spells Completed in 9 to 12 Months
Employer-sponsored 35 10 2.4 0 47.4
Medicaid 16.2 26 1.1 0.5 43.9
Other 3.3 1.7 0.8 0 5.8
Not in Universe 0 3 0 0 3
Total 54.5 40.7 4.3 0.5 100
  Spells Completed in 13 to 18 Months*
Employer-sponsored 21.8 18.6 8.2 3.5 52.1
Medicaid 9.7 38.2 0 0 47.9
Other 0 0 0 0 0
Not in Universe 0 0 0 0 0
Total 31.5 56.8 8.2 3.5 100

* These spells are limited to those beginning in the first six months of FY93.

SOURCE: Survey of Income and Program Participation, 1992 Panel.

The estimates reported in each subtable are percentages of the total number of spells in each duration group. Thus the percentages in the total column at the right give the distribution of insurance coverage prior to the spell of uninsurance; the percentages in the total row beneath each subtable give the distribution of insurance coverage after the spell of uninsurance; and the percentages inside the row and column margins give the distribution of spells by combinations of prior and subsequent insurance coverage.

For all spells of uninsurance completed in 12 months or less, 54 percent were preceded by employer-sponsored coverage, 39 percent were preceded by Medicaid, 4.5 percent were preceded by another form of coverage, and 2.5 percent were experienced by children who were not in the survey universe prior to the spell. These last would be newborns who were without health insurance at birth. The distribution of sources of coverage following a spell of uninsurance is very similar to the distribution prior to the spell, with 53 percent being covered by an employer-sponsored plan, 40 percent by Medicaid, 5 percent by another type of plan, and 1.6 percent leaving the survey universe.6

In October 1992, based on these same data, about 19 percent of insured children were covered by Medicaid, 76 percent by an employer-sponsored plan, and 5 percent by another type of plan. Thus the frequency of Medicaid as the source of coverage before and after a spell of uninsurance lasting 12 months or less is more than double the frequency of Medicaid among all insured children. Coverage other than Medicaid or an employer-sponsored plan occurs with about the same frequency among all insured children as among those who later or previously experienced spells of uninsurance, while employer-sponsored coverage occurs with lower frequency before or after a spell of uninsurance than it does among all insured children.

The similarity of the distributions of sources of coverage before and after spells of uninsurance does not necessarily mean that children end their spells of uninsurance by returning to the same coverage that they lost to begin their spells, but we see from the cross-tabulation of coverage before and after uninsurance that the two sources are clearly not independent. For all spells of uninsurance completed in 12 months or less, 28 percent were preceded and followed by Medicaid enrollment, and 42 percent were preceded and followed by employer-sponsored coverage. About 9 percent of children changed from employer-sponsored coverage to Medicaid, and somewhat fewer children made the reverse transition.

Among spells completed in 12 months or less, we see little variation in these basic patterns by duration. But spells completed in 13 to 18 months present a very different picture, showing an even greater prominence of Medicaid. Prior to a spell lasting 13 to 18 months, 48 percent of the children were insured by Medicaid and 52 percent by an employer-sponsored plan. We found no incidence of other sources of coverage and no children who remained uninsured for 13 months past their births. Following a spell of 13 to 18 months, 57 percent of children enrolled in Medicaid compared to 32 percent becoming covered by an employer-sponsored plan. Medicaid was the source of coverage both before and after a long spell of uninsurance among 38 percent of the children who experienced such spells. Employer-sponsored coverage preceded and followed 22 percent of these spells--barely half the percentage among shorter spells. Children who left a spell of uninsurance that was preceded by employer-sponsored coverage were almost as likely to end their spells by acquiring Medicaid coverage as employer-sponsored coverage, and other coverage was nearly half as common as Medicaid. A possible explanation for the differential importance of Medicaid by duration is that parents who lose employer-sponsored coverage tend to defer pursuing other coverage for their children because they expect to regain employer-sponsored coverage. If they are unable to regain employer-sponsored coverage within a year or so, they try to obtain Medicaid. It could be, as well, that many of these children did not qualify for Medicaid in the months after employer-sponsored coverage was lost and became eligible only after their parents spent down savings and other resources.

2. Spells of Medicaid-Eligible Uninsurance

A spell of uninsurance in which a child is eligible for Medicaid can begin or end with the child losing Medicaid eligibility without regaining insurance coverage. Table 6 presents cross-tabulations of coverage before and after spells of Medicaid-eligible uninsurance, but it differs from Table 5 in that it includes uninsurance without Medicaid eligibility as a status before and after a spell.

Nearly one-half of children who experienced spells of Medicaid-eligible uninsurance lasting 12 months or less were uninsured but ineligible for Medicaid before beginning their spells, and 31 percent left spells of Medicaid-eligible uninsurance by losing their Medicaid eligibility. Medicaid enrollment preceded 27 percent of the spells of Medicaid-eligible uninsurance and followed 31 percent of them. Only 19 percent of children who experienced spells of Medicaid-eligible uninsurance were covered by employer-sponsored insurance immediately prior to their spells, and just under 15 percent gained employer-sponsored coverage to end their spells. Uninsurance without Medicaid-eligibility bracketed 34 percent of the 1 to 12 month spells of Medicaid-eligible uninsurance while actual Medicaid enrollment bracketed 16 percent of the spells. Employer- sponsored coverage preceded and followed only 6 percent of spells.

As we have seen, spells of Medicaid-eligible uninsurance tend to be quite short, so the spells completed in 12 months or less are dominated by spells completed in 1 to 4 months. The pattern of coverage before and after spells of Medicaid-eligible uninsurance changes dramatically as the duration of these spells increases. Until durations of 13 months or longer the changes are largely confined to the distribution of coverage prior to the spell of Medicaid-eligible uninsurance. Medicaid enrollment becomes more important than Medicaid ineligibility, accounting for more and more of the coverage prior to such spells. Among spells completed in 13 to 18 months, 43 percent began with children leaving Medicaid enrollment compared to 31 percent that started with uninsured children becoming Medicaid-eligible. The reverse trend occurred for the source of coverage following a spell of Medicaid-eligible uninsurance, although there are minimal changes between spells of 1 to 4 month duration and spells of 9 to 12 month duration. Among the longer spells, 65 percent ended with children losing Medicaid eligibility but remaining uninsured. Only 6 percent of the children gained employer-sponsored coverage while 29 percent enrolled in Medicaid.

TABLE 6: INSURANCE COVERAGE BEFORE AND AFTER COMPLETED SPELLS OF MEDICAID- ELIGIBLE UNINSURANCE BEGINNING IN FY93, BY DURATION OF SPELL
  Insurance Coverage After Completion of Spell
Insurance Coverage Before Start of Spell Employer- sponsored Medicaid Other Uninsured (Med-inel.) Not in Universe
  All Spells Completed in 12 Months or Less
Employer-sponsored 6 4.8 0.2 8.1 0.2
Medicaid 2.4 15.6 0.4 8.4 0.6
Other 0.3 0.4 0 0.6 0
Uninsured (Med-inel.) 5.6 9.2 0.4 34.4 0
Not in Universe 0.4 1.5 0 0.5 0.1
Total 14.7 31.4 1 51.9 0.9
  Spells Completed in 1 to 4 Months
Employer-sponsored 6.3 4.7 0.3 7 0.2
Medicaid 2.1 15 0.2 6.8 0.7
Other 0.1 0.3 0 0.7 0
Uninsured (Med-inel.) 6.1 9.6 0.5 37.5 0
Not in Universe 0.2 1 0 0.6 0.2
Total 14.8 30.5 1 52.6 1.1
  Spells Completed in 5 to 8 Months
Employer-sponsored 5.6 5.4 0 12 0
Medicaid 1.5 19.6 0.7 11.2 0
Other 1.3 0 0 0.6 0
Uninsured (Med-inel.) 6 7.2 0 24.3 0
Not in Universe 1.7 2.9 0 0 0
Total 16.1 35.1 0.7 48.1 0
  Spells Completed in 9 to12 Months
Employer-sponsored 3.3 5.4 0 13.1 0
Medicaid 7.2 14.2 2.1 21.5 0.9
Other 0 2 0 0 0
Uninsured (Med-inel.) 0 8.9 0 17.1 0
Not in Universe 0 4.1 0 0 0
Total 10.6 34.6 2.1 51.8 0.9
  Spells Completed in 13 to 18 Months*
Employer-sponsored 0 0 0 21.1 0
Medicaid 5.9 28.8 0 8.8 0
Other 0 0 0 0 0
Uninsured (Med-inel.) 0 0 0 31.2 0
Not in Universe 0 0 0 4.3 0
Total 5.9 28.8 0 65.3 0

* These spells are limited to those beginning in the first six months of FY93.

SOURCE: Survey of Income and Program Participation, 1992 Panel.

In looking at the combinations of coverage that precede and follow 13 to 18 month spells of Medicaid-eligible uninsurance, we find that 29 percent of these spells are bracketed by Medicaid enrollment and 31 percent are bracketed by uninsurance without Medicaid eligibility. We found no instances of children having employer-sponsored coverage both before and after a spell of Medicaid- eligible uninsurance. All of the children who lost employer-sponsored coverage to become uninsured but eligible for Medicaid ended their spells by losing their Medicaid eligibility rather than enrolling in Medicaid. This is quite different from what we observed with all spells of uninsurance, but we cannot tell from Table 6 what may have happened to these children after losing Medicaid eligibility. Many of them may have gained employer-sponsored coverage, and some could have regained Medicaid eligibility and become enrolled. Finally, in viewing the bottom panel of Table 6 we need to keep in mind that, according to Table 4, these spells account for less than four percent of all spells of Medicaid-eligible uninsurance that began in FY93.

D. Transitions

In the previous section we examined transitions into and out of uninsurance but restricted our attention to completed spells of uninsurance and, even more narrowly, completed spells of Medicaid- eligible uninsurance.. In this section we address the issue of transitions more broadly, looking first at changes in health insurance coverage at roughly one- and two-year intervals, then month-to-month transitions and, finally, cumulative monthly exits from uninsurance and Medicaid enrollment.

1. Change in Coverage between Years

Table 7 provides a cross-tabulation of health insurance coverage in October 1992 by health insurance coverage in September 1993 and in September 1994. The top row reports the number of children who were in each coverage group in October 1992. Below that is reported, for each October 1992 coverage group, the percentage distribution of children in September 1993 and, in the lower panel, September 1994.

Of the 12.6 million children who were reported as enrolled in Medicaid in October 1992, about 80 percent were still enrolled in September 1993 and 74 percent in September 1994. These figures do not reflect continuous enrollment, necessarily. A child who left Medicaid during 1993 but returned by September would be included in the 80 percent. Nevertheless, they indicate a very high degree of continuity in the reliance on Medicaid. Most of those who were not enrolled in Medicaid one or two years later were uninsured: 11 percent in September 1993 and nearly 14 percent in September 1994. Most of the remainder--7 percent in September 1993 and close to 10 percent in September 1994--acquired employer-sponsored coverage.

Of the 9.5 million children who were uninsured in October 1992, 62 percent were still uninsured in September 1993 and 52 percent in September 1994. Those who became insured were more likely to find such coverage in employer-sponsored plans than in Medicaid but at nowhere near the four to one rate that characterizes the dominance of employer-sponsored coverage over Medicaid among all insured children. In September 1993, 20 percent of the children who were uninsured 11 months earlier had employer-sponsored coverage while 14 percent were enrolled in Medicaid. One year later, 25 percent had employer-sponsored coverage, and 17 percent were covered by Medicaid. Other coverage accounted for 2.5 percent of these children in September 1993 and 3.7 percent in September 1994 while 2.1 percent had left the SIPP population by the first date and 3.2 percent by the second. In September 1993 we also find that 4.4 percent of those with employer-sponsored coverage in October 1992 had become uninsured compared to 8.6 percent of those with other coverage. By September 1994 the percentage uninsured among those with employer-sponsored coverage in October 1992 had risen modestly to 6.1 percent while the percentage uninsured among those with other coverage initially had actually fallen slightly.

TABLE 7: DISTRIBUTION OF CHILDREN BY HEALTH INSURANCE COVERAGE IN OCTOBER 1992 AND COVERAGE IN SEPTEMBER 1993 AND SEPTEMBER 1994
    Health Insurance Coverage in October 1992
Calendar Month and Coverage Medicaid Employer- Sponsored Coverage Other Coverage Uninsured Total
October 1992--Total Children 12,614,000 46,439,000 3,007,000 9,489,000 71,549,000
    Coverage in September 1993          
Medicaid 79.8 2.5 3.3 13.9 17.7
Employer-sponsored Coverage 7.1 90.8 29.9 19.7 64
Other Coverage 0.9 1.7 57.8 2.5 4
Uninsured 11.4 4.4 8.6 61.8 13.4
Not in Population 0.8 0.5 0.4 2.1 0.8
    Total 100 100 100 100 100
    Coverage in September 1994          
Medicaid 74.1 2.7 2.9 16.8 17.2
Employer-sponsored Coverage 9.6 88 36.9 24.8 63.6
Other Coverage 1 2.3 52.6 3.7 4.3
Uninsured 13.6 6.1 7.1 51.5 13.4
Not in Population 1.7 1.1 0.5 3.2 1.4
Total 100 100 100 100 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

2. Month-to-Month Transitions

Changes in the health insurance coverage of individuals between points in time one and two years apart give us a very good idea of the magnitudes of gross flows of children into and out of particular types of coverage over the span of one and two years. But they do not show all movement. In particular, they do not capture multiple changes by the same individuals. To gain a different perspective on changes in health insurance coverage over time, we examine changes between one month and the next. More specifically, we examine month-to-month transitions into and out of uninsurance and into and out of Medicaid.

a. Uninsurance

How many children leave uninsurance each month and how many enter uninsurance each month? Tables 8 and 9 present estimates of monthly transitions out of and into uninsurance, respectively. Each of these tables is based on 23 successive cross-tabulations of insurance status between one month and the next.

TABLE 8: CHILDREN UNDER 19 WITHOUT HEALTH INSURANCE BY INSURANCE STATUS IN NEXT MONTH: FY 1993 and 1994
    Type of Coverage in Next Month   Type of Coverage in Next Month
Next Month among Children
Who Become Insured
Current Month Number Without Insurance Uninsured Medicaid Employer Other Out of Universe Number
Becoming
Insured
Medicaid Employer Other
9210 9,489,000 90.5% 4.2% 4.0% 0.4% 1.0% 809,000 48.8% 46.4% 4.8%
9211 9,238,000 91.0% 3.6% 4.1% 0.5% 0.8% 749,000 44.3% 50.0% 5.7%
9212 9,141,000 91.1% 2.6% 4.8% 0.7% 0.7% 745,000 31.8% 59.3% 8.9%
9301 9,126,000 92.1% 3.3% 3.5% 0.5% 0.6% 672,000 44.9% 48.1% 6.9%
9302 9,144,000 91.0% 3.3% 4.9% 0.3% 0.4% 781,000 39.1% 57.8% 3.2%
9303 8,974,000 93.1% 2.6% 3.2% 0.5% 0.6% 564,000 39.1% 50.2% 7.8%
9304 9,133,000 91.0% 3.1% 5.5% 0.2% 0.1% 804,000 35.4% 62.8% 1.8%
9305 9,137,000 89.7% 4.7% 4.1% 0.8% 0.7% 876,000 49.3% 42.3% 8.4%
9306 8,947,000 92.1% 2.9% 3.9% 0.1% 1.0% 614,000 42.0% 56.3% 1.7%
9307 9,222,000 90.0% 3.7% 4.5% 0.8% 0.9% 835,000 40.5% 50.2% 9.3%
9308 9,134,000 90.5% 4.3% 3.8% 0.5% 0.9% 791,000 49.6% 44.1% 6.3%
9309 9,271,000 90.4% 3.6% 4.8% 0.7% 0.4% 849,000 39.9% 52.7% 7.4%
9310 9,080,000 91.3% 3.6% 4.1% 0.5% 0.4% 752,000 44.0% 49.8% 6.2%
9311 8,920,000 92.5% 2.5% 4.0% 0.2% 0.8% 598,000 37.6% 59.7% 2.8%
9312 9,021,000 90.6% 3.5% 4.3% 0.6% 1.0% 751,000 41.9% 51.1% 7.0%
9401 9,009,000 90.9% 3.0% 4.9% 0.6% 0.6% 767,000 35.5% 58.0% 6.5%
9402 9,098,000 91.4% 2.9% 4.1% 1.0% 0.6% 725,000 36.4% 51.0% 12.7%
9403 9,050,000 91.2% 3.6% 4.0% 0.6% 0.6% 744,000 43.5% 49.1% 7.4%
9404 8,998,000 92.8% 2.2% 3.8% 0.6% 0.6% 596,000 32.8% 57.7% 9.5%
9405 9,025,000 89.6% 4.4% 4.5% 0.8% 0.6% 884,000 45.4% 46.1% 8.4%
9406 8,719,000 89.9% 4.3% 4.0% 1.0% 0.7% 816,000 46.3% 42.9% 10.8%
9407 8,694,000 89.5% 3.5% 5.0% 0.5% 1.5% 782,000 38.5% 55.6% 5.9%
9408 8,745,000 90.3% 2.3% 5.7% 1.1% 0.7% 790,000 24.9% 63.4% 11.7%
Average 9,057,000 91.0% 3.4% 4.3% 0.6% 0.7% 752,000 40.8% 52.1% 7.1%
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 9: CHILDREN UNDER 19 WITHOUT HEALTH INSURANCE BY INSURANCE STATUS IN PRECEDING MONTH: FY 1993 and 1994
    Type of Coverage in Preceding Month   Insurance Status in Preceding Month Among Children Who Become Uninsured
Current Month Number Without Insurance Uninsured Medicaid Employer Other Out of Universe Number Becoming Uninsured Medicaid Employer Other Not in Universe
9211 9,238,000 93.0% 2.3% 4.0% 0.4% 0.3% 650,000 32.1% 57.4% 6.1% 4.5%
9212 9,141,000 92.0% 2.8% 4.5% 0.5% 0.2% 731,000 34.7% 56.8% 6.1% 2.3%
9301 9,126,000 91.3% 3.3% 4.6% 0.5% 0.3% 796,000 38.4% 52.9% 5.3% 3.4%
9302 9,144,000 91.9% 3.6% 3.7% 0.4% 0.4% 742,000 43.9% 45.9% 5.1% 5.1%
9303 8,974,000 92.8% 2.3% 4.4% 0.4% 0.2% 649,000 32.1% 60.5% 4.9% 2.4%
9304 9,133,000 91.5% 3.6% 4.2% 0.4% 0.3% 775,000 42.1% 49.8% 4.9% 3.1%
9305 9,137,000 91.0% 4.3% 4.2% 0.3% 0.2% 822,000 48.2% 46.6% 3.5% 1.7%
9306 8,947,000 91.6% 2.8% 4.5% 0.8% 0.2% 748,000 33.4% 54.0% 9.7% 2.8%
9307 9,222,000 89.3% 4.0% 5.5% 1.0% 0.1% 982,000 37.5% 51.4% 9.8% 1.3%
9308 9,134,000 90.9% 3.1% 5.1% 0.5% 0.4% 830,000 33.8% 56.3% 5.7% 4.2%
9309 9,271,000 89.1% 4.5% 5.9% 0.2% 0.3% 1,006,000 41.0% 53.9% 2.0% 3.0%
9310 9,080,000 92.3% 4.2% 2.6% 0.4% 0.4% 697,000 54.9% 34.4% 5.5% 5.3%
9311 8,920,000 92.9% 3.2% 3.7% 0.0% 0.2% 631,000 45.5% 52.0% 0.0% 2.6%
9312 9,021,000 91.5% 2.8% 5.2% 0.2% 0.3% 770,000 33.0% 61.1% 2.1% 3.9%
9401 9,009,000 90.7% 3.6% 4.7% 0.6% 0.4% 834,000 39.0% 50.9% 6.0% 4.2%
9402 9,098,000 90.0% 5.1% 3.9% 0.9% 0.1% 908,000 51.0% 38.9% 8.7% 1.4%
9403 9,050,000 91.9% 3.3% 4.4% 0.4% 0.1% 735,000 40.6% 53.6% 5.1% 0.8%
9404 8,998,000 91.7% 2.9% 4.8% 0.5% 0.2% 745,000 34.9% 57.4% 5.8% 1.9%
9405 9,025,000 92.5% 3.4% 3.0% 0.9% 0.2% 675,000 45.5% 39.6% 12.3% 2.7%
9406 8,719,000 92.8% 2.7% 4.0% 0.4% 0.2% 631,000 36.8% 55.2% 5.9% 2.1%
9407 8,694,000 90.2% 3.6% 5.5% 0.7% 0.1% 853,000 36.4% 55.8% 6.7% 1.1%
9408 8,745,000 88.9% 3.9% 5.8% 0.9% 0.4% 967,000 35.1% 52.7% 8.2% 4.0%
9409 8,911,000 88.6% 3.9% 6.9% 0.2% 0.4% 1,017,000 34.4% 60.2% 1.8% 3.6%
Average 9,032,000 91.2% 3.4% 4.6% 0.5% 0.3% 791,000 39.3% 52.1% 5.7% 2.9%
SOURCE: Survey of Income and Program Participation, 1992 Panel.

For each month from October 1992 (9210) through August 1994 (9408), Table 8 reports the total number of children without health insurance, followed by a percentage distribution of their coverage in the next month, the number who became insured, and a percentage distribution of the type of insurance coverage that these latter acquired. Average monthly estimates of each of these quantities are presented at the bottom of the table, and we consider these first. On average, 9 percent of the roughly 9 million uninsured children left that state each month, with 8.3 percent becoming insured and the remainder leaving the survey universe or aging out of the child population.7 Of the 752,000 who became insured each month, 41 percent enrolled in Medicaid, 52 percent gained coverage under an employer-sponsored plan, and 7 percent obtained coverage by another type of insurance.8

The 752,000 children, on average, who became insured between one month and the next plus the additional 64,000 children who left the survey universe or turned 19 years old each month represent a sizable number. Outflows of this magnitude would have depleted the stock of uninsured children in a little over a year had there been no flows in the opposite direction. Of course, there were flows in the opposite direction, and they have to compare in magnitude to the flows out of uninsurance as the net change in the number of uninsured children over the 23 months is less than the outflow in a single month. Moreover, we know from Table 7 that 62 percent of the children who were uninsured in October 1992 were still uninsured in September 1993, and nearly 52 percent were still uninsured a year later. From the relative frequency of short durations among new spells of uninsurance, reported earlier, we can infer that most of the children who exited uninsurance during FY93 and FY94 did not become uninsured until after the start of the period. It is because of the number of children flowing into uninsurance, as we shall see below, that 9 percent of uninsured children could leave that population month after month with little effect on the total number of children who were uninsured at a point in time.

The individual monthly results that make up the bulk of Table 8 are of interest for what information they may provide on seasonal patterns and trends in exits from uninsurance. In research conducted for the Food and Nutrition Service, using the 1992 and 1993 SIPP panels, Czajka et al. (1998) found clear evidence of seasonality in the total number of children in poverty. In the research presented in Technical Appendix A, however, we found no evidence of seasonality in the total number of uninsured children. If there are any seasonal patterns in the movement of children into and out of uninsurance, of course, they would show up most clearly in the numbers of children entering and leaving uninsurance each month and in the distribution of their sources of coverage before and after spells of uninsurance.

We see no evidence of seasonality in any aspect of the monthly transitions reported in Table 8. Nor do we see clear evidence of a trend in the number becoming uninsured, even though the number without insurance declines by more than 700,000 over the 23 months. While the monthly estimates of children leaving uninsurance bounce around quite a lot, varying from a low of 564,000 to a high of 884,000, this variation appears to be due entirely to sampling error.

Table 9 reports estimates of the number of children who were without health insurance in November 1992 through September 1994, the percentage distribution of their coverage in the preceding month, the number who became uninsured in the current month, and the percentage distribution of insurance status in the preceding month among those who became uninsured. On average, 791,000 children became uninsured each month. Of this number, 52 percent had employer coverage in the preceding month, 39 percent had Medicaid coverage, about 6 percent had other coverage, and less than 3 percent were newborns or children returning to the survey universe. The distribution of prior month coverage among those becoming uninsured compares fairly closely to the distribution of prior month coverage among children who began spells of uninsurance in FY93 that lasted one to 12 months (see Table 5). As we explained above, with reference to the distribution of coverage after completion of a spell of uninsurance, this similarity of the two distributions is not at all surprising, given that Table 8 refers to all new spells that started in FY93 or FY94 (exclusive of October 1992) and Table 5 refers to a large subset of the spells that began in FY93.

In the monthly results the numbers who became uninsured in September 1993 and 1994 stand out because they are the only estimates in excess of one million. Nevertheless, there is no other evidence of seasonal change in the monthly estimates, and so we infer that there is no seasonality in the movement of children into uninsurance. Similarly, we see no seasonal swings in any of the components of change. These vary a lot from month to month but with no clear pattern, and so we conclude that their variation is due to sampling error, essentially. Nor do there appear to be any long-term trends in either the number becoming uninsured or the composition of coverage in the preceding month. The final months show a rise in the number becoming uninsured and, with it, the proportion of uninsured children who were insured the preceding month, but these are strictly short- term changes that in reality may represent nothing more than sampling error.

b. Medicaid

Table 10 shows for each month from October 1992 through August 1994 the number of children under 19 who were enrolled in Medicaid, a percentage breakdown of the types of coverage they had in the next month, the number of children leaving Medicaid in the next month, and a percentage distribution of the type of coverage that followed Medicaid. Average monthly transitions are reported at the bottom of the table. On average about 4 percent of Medicaid enrollees under 19 years of age or more than half a million children (555,000) left Medicaid each month. Of these, 56 percent became uninsured while about 37 percent acquired another form of insurance and 7 percent left the survey universe or aged out of the population of children under 19.

Table 11 reports monthly transitions into Medicaid. Of the 583,000 entering Medicaid each month, on average, just over half or 53 percent were previously uninsured, 34 percent had employer- sponsored or other insurance coverage, and 14 percent were either new to the survey universe or returning to the survey universe. In fact, virtually all if not all of those entering the survey universe were newborn children who were enrolled in Medicaid at birth (an extension of their mothers’ coverage in pregnancy, no doubt). The size of this group reflects both the high rate of Medicaid coverage among infants and the disproportionately large share of the Medicaid population that infants represent.

TABLE 10: CHILDREN UNDER 19 ENROLLED IN MEDICAID BY INSURANCE STATUS IN NEXT MONTH: FY 1993 and 1994
    Type of Coverage in Next Month   Type of Coverage in Next Month Among Children Who Leave  Medicaid
Current Month Number Enrolled in Medicaid Medicaid Employer Other Uninsured Out of Universe Number Leaving Medicaid Employer Other Uninsured Out of Universe
9210 12614000 96.3% 1.7% 0.0% 1.7% 0.3% 468000 47.1% 0.0% 44.6% 8.3%
9211 12800000 95.8% 2.0% 0.1% 2.0% 0.2% 541000 46.2% 1.2% 46.9% 5.6%
9212 12829000 96.1% 1.1% 0.0% 2.4% 0.4% 503000 28.8% 1.1% 60.8% 9.4%
9301 12842000 95.7% 1.4% 0.2% 2.5% 0.3% 558000 31.8% 3.5% 58.3% 6.4%
9302 12901000 97.1% 0.9% 0.1% 1.6% 0.3% 374000 32.0% 2.7% 55.8% 9.5%
9303 13187000 95.9% 1.4% 0.0% 2.5% 0.2% 539000 35.0% 0.0% 60.6% 4.4%
9304 13264000 94.5% 1.8% 0.2% 3.0% 0.5% 735000 33.2% 3.0% 54.0% 9.8%
9305 12983000 96.3% 1.3% 0.1% 1.9% 0.4% 474000 35.3% 2.3% 52.7% 9.7%
9306 13328000 95.4% 1.3% 0.2% 2.8% 0.4% 618000 28.2% 4.5% 59.6% 7.7%
9307 13294000 95.5% 1.8% 0.1% 2.1% 0.5% 600000 39.6% 2.5% 46.8% 11.1%
9308 13233000 95.6% 0.7% 0.0% 3.1% 0.5% 584000 16.3% 1.0% 70.7% 12.0%
9309 13369000 94.8% 2.0% 0.1% 2.9% 0.3% 701000 38.2% 1.9% 54.6% 5.3%
9310 13236000 95.9% 1.9% 0.0% 2.2% 0.1% 541000 45.6% 0.0% 53.1% 1.3%
9311 13263000 95.7% 1.9% 0.4% 1.9% 0.1% 569000 43.5% 8.5% 44.6% 3.4%
9312 13233000 95.5% 1.6% 0.2% 2.5% 0.3% 595000 35.0% 3.5% 54.6% 7.0%
9401 13231000 95.4% 0.8% 0.1% 3.5% 0.1% 605000 18.2% 2.1% 76.5% 3.2%
9402 13228000 95.8% 1.1% 0.3% 2.3% 0.6% 560000 25.1% 7.9% 53.3% 13.8%
9403 13079000 96.1% 1.6% 0.1% 2.0% 0.2% 514000 40.5% 3.0% 50.6% 5.9%
9404 13185000 96.2% 1.1% 0.0% 2.3% 0.4% 505000 28.6% 1.2% 60.8% 9.4%
9405 13109000 96.7% 1.4% 0.0% 1.8% 0.1% 432000 43.1% 0.0% 53.7% 3.2%
9406 13270000 96.0% 1.4% 0.0% 2.3% 0.3% 526000 34.2% 0.0% 59.0% 6.9%
9407 13372000 95.1% 1.7% 0.0% 2.5% 0.6% 652000 35.1% 0.0% 52.1% 12.8%
9408 13418000 95.7% 1.6% 0.1% 2.6% 0.0% 577000 37.3% 2.1% 60.6% 0.0%
Average 13142000 95.8% 1.5% 0.1% 2.4% 0.3% 555000 34.5% 2.3% 56.0% 7.3%
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 11: CHILDREN UNDER 19 ENROLLED IN MEDICAID BY INSURANCE STATUS IN PRECEDING MONTH: FY 1993 and 1994
 
    Type of Coverage in Preceding Month   Type of Coverage in Preceding Month Among Children Who Enter Medicaid
Current Month Number Enrolled in Medicaid Medicaid Employer Other Uninsured Out of Universe Number Entering Medicaid Employer Other Uninsured Not in Universe
9211 12800000 94.9% 1.1% 0.1% 3.1% 0.8% 655000 21.0% 2.8% 60.3% 15.9%
9212 12829000 95.6% 1.3% 0.0% 2.6% 0.6% 570000 29.1% 0.0% 58.3% 12.5%
9301 12842000 96.0% 1.3% 0.5% 1.8% 0.4% 516000 31.5% 12.7% 45.8% 10.0%
9302 12901000 95.2% 1.4% 0.2% 2.3% 0.8% 617000 30.0% 3.9% 49.0% 17.1%
9303 13187000 95.0% 1.6% 0.0% 2.3% 1.1% 660000 32.0% 0.0% 46.2% 21.8%
9304 13264000 95.4% 1.7% 0.1% 1.8% 1.1% 616000 36.0% 2.5% 38.4% 23.0%
9305 12983000 96.5% 0.7% 0.1% 2.2% 0.6% 453000 18.8% 2.5% 62.8% 15.9%
9306 13328000 93.9% 2.1% 0.0% 3.2% 0.8% 819000 33.5% 0.0% 52.7% 13.8%
9307 13294000 95.6% 1.7% 0.0% 1.9% 0.8% 584000 38.0% 0.0% 44.2% 17.8%
9308 13233000 95.9% 1.0% 0.0% 2.6% 0.5% 540000 25.1% 0.0% 62.7% 12.2%
9309 13369000 94.6% 1.9% 0.0% 2.9% 0.6% 720000 34.7% 0.0% 54.5% 10.7%
9310 13236000 95.7% 1.3% 0.0% 2.6% 0.4% 567000 30.2% 0.9% 59.6% 9.3%
9311 13263000 95.7% 1.1% 0.1% 2.5% 0.6% 568000 25.5% 3.1% 58.3% 13.1%
9312 13233000 95.9% 1.7% 0.2% 1.7% 0.5% 540000 41.6% 3.7% 41.6% 13.0%
9401 13231000 95.5% 1.7% 0.0% 2.4% 0.4% 593000 38.5% 0.0% 53.1% 8.4%
9402 13228000 95.4% 2.0% 0.0% 2.1% 0.5% 602000 43.5% 1.1% 45.1% 10.3%
9403 13079000 96.9% 0.3% 0.1% 2.0% 0.7% 412000 9.4% 4.4% 64.1% 22.2%
9404 13185000 95.3% 1.6% 0.1% 2.5% 0.5% 621000 34.4% 2.9% 52.1% 10.6%
9405 13109000 96.7% 1.4% 0.0% 1.5% 0.4% 428000 41.5% 0.0% 45.7% 12.8%
9406 13270000 95.5% 0.9% 0.2% 3.0% 0.3% 594000 20.2% 4.7% 67.6% 7.5%
9407 13372000 95.3% 1.5% 0.0% 2.8% 0.4% 628000 32.0% 0.0% 60.1% 7.9%
9408 13418000 94.8% 1.9% 0.0% 2.2% 1.0% 698000 36.1% 0.9% 43.2% 19.8%
9409 13259000 96.8% 1.2% 0.0% 1.5% 0.5% 418000 37.9% 0.0% 47.1% 14.9%
Average 13170000 95.6% 1.4% 0.1% 2.3% 0.6% 583000 31.6% 1.9% 52.6% 13.9%
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Neither Table 10 nor Table 11 shows any evidence of seasonality or trend in the exits from or entrances into Medicaid enrollment. That Medicaid enrollment among the SIPP population of children under 19 grows by nearly 500,000 over the period can be attributed to the fact that new enrollments tended to outnumber exits from Medicaid in each month rather than one series growing or declining more rapidly than the other.

3. Cumulative Monthly Exits

The monthly transitions reported in Table 8 show that about 8 percent of the children who were uninsured at any point in time in FY93 or FY94 became insured in the next month while Table 7 indicates that 36 percent of the children who were uninsured in October 1992 were insured 11 months later and 46 percent were insured 23 months later. Similarly, the monthly transitions reported in Table 10 show that, on average, about 4 percent of the children enrolled in Medicaid at any time in the two-year period left the program the next month while Table 7 shows that about 20 percent exited over the course of 11 months and 26 percent over 23 months. In each case the change over 11 or 23 months is considerably less than we would see if each cohort of children who were uninsured or enrolled in Medicaid in a given month experienced the same exit rates month after month.9 Here we examine the cumulative effect of monthly transitions on children who started out in a particular state of health insurance coverage in October 1992. For each of three initial statuses-- uninsurance, Medicaid-eligible uninsurance, and Medicaid enrollment--we present estimates of the distribution of children by coverage status in each of the next 23 months. Over time, then, we see what fraction of the original population of, say, uninsured children is still uninsured in each of the next 23 months and what fraction has moved to either Medicaid, employer-sponsored insurance, or other insurance.

a. Uninsurance

Table 12 presents the distribution of children who were uninsured in October 1992 by their health insurance coverage in each of the next 23 months.(10) As we saw in Table 8, about 9 percent leave the uninsured population in the first month. This exit rate begins declining immediately, dropping to between 2 and 3 percent within four months and to about 1 percent within another four months. By the final months of FY94, the percentage remaining uninsured is declining by less than one percentage point per month (or less than 2 percent of those remaining uninsured). Initially, children who leave uninsurance are about equally likely to enroll in Medicaid and obtain employer- sponsored coverage, but after two months those children who have left uninsurance are increasingly more likely to be covered by employer-sponsored coverage than by Medicaid. For some of the children who end up with employer-sponsored coverage, the path from uninsurance to employer- sponsored coverage may pass through Medicaid. That is, children may leave uninsurance by obtaining Medicaid coverage but then move from Medicaid to employer-sponsored coverage at a later date. We cannot tell from this table the paths that children take. On the other hand, it is quite clear from these data that as long as 23 months after children have left uninsurance, Medicaid remains a much more important source of coverage than it is among all insured children. In September 1994, children who were uninsured in October 1992 are less than one-and-a-half times

TABLE 12: DISTRIBUTION OF CHILDREN WHO WERE UNINSURED IN OCTOBER 1992
BY HEALTH INSURANCE COVERAGE IN EACH OF NEXT 23 MONTHS
  Health Insurance Coverage in Calendar Month
Reference Month and Calendar Month
 
Uninsured Medicaid Employer- Sponsored Coverage Other Coverage Not in Population Total
0. October 1992 100 0 0 0 0 100
1. November 1992 90.9 4.3 4 0.4 0.5 100
2. December 1992 83.3 7.4 7.9 0.9 0.5 100
3. January 1993 76.3 9.5 11.5 1.7 1 100
4. February 1993 71.4 11.8 13.7 2 1.1 100
5. March 1993 68.8 12.2 15.8 2 1.1 100
6. April 1993 68.4 11.9 16.3 2 1.4 100
7. May 1993 65.8 12.7 18 2 1.5 100
8. June 1993 63.7 13.6 19 2 1.8 100
9. July 1993 64.4 13.5 18.2 2 2 100
10. August 1993 62.7 13.2 20 2.2 2 100
11. September 1993 61.8 13.9 19.7 2.5 2.1 100
12. October 1993 59.6 14.7 20.5 3 2.2 100
13. November 1993 58.9 14.8 20.9 3.3 2.1 100
14. December 1993 59 14.8 20.8 3.1 2.3 100
15. January 1994 58.3 15.2 21 3.2 2.3 100
16. February 1994 58.6 14.4 21.7 2.9 2.4 100
17. March 1994 58.1 14.5 22.2 2.8 2.4 100
18. April 1994 55.9 16.3 22.1 3.3 2.4 100
19. May 1994 55.6 16.1 22.5 3.4 2.5 100
20. June 1994 53.1 17.1 23.4 3.8 2.6 100
21. July 1994 52 17.9 22.9 4.3 3 100
22. August 1994 52.1 17.7 23.6 3.4 3.1 100
23. September 1994 51.5 16.8 24.8 3.7 3.2 100

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: The number of children under 19 who were uninsured in October 1992 is 9,489,000.

as likely to have employer-sponsored coverage as Medicaid (24.8 percent versus 16.8 percent of all children who were uninsured in October 1992). Among all insured children in September 1994 who were under 19 and part of the SIPP universe in October 1992, the number covered by employer- sponsored plans is nearly four times the number covered by Medicaid (45.5 million versus 12.3 million).

b. Medicaid-eligible Uninsurance

Table 13 reports the distribution of children who were uninsured and eligible for Medicaid in October 1992 by their health insurance coverage in each of the next 23 months. Within one month, more than 20 percent of the children who were Medicaid-eligible and uninsured had changed to another coverage status. About 7 percent remained uninsured but were no longer eligible for Medicaid while a comparable percentage had enrolled in Medicaid. Just over 5 percent had employer-sponsored coverage while only .2 percent had other coverage. About .6 percent had left the population. After two months, the percentage remaining Medicaid-eligible and uninsured had dropped to under 65 percent, and within seven or eight months it had reached 40 percent. After 23 months, only 26 percent of the original population of children who were Medicaid-eligible and uninsured at the outset remained part of that population. Note, however, that 21 percent of the original population were still uninsured. Comparing Tables 13 and 12, we see that the Medicaid- eligible uninsured were only slightly more likely to have left the state of uninsurance than all uninsured: 52.5 percent versus 48.5 percent. Even so, Medicaid remained the dominant source of coverage among those who left uninsurance, accounting for 27 percent of the original population compared to 18 percent with employer-sponsored coverage. This represents a reversal of the relative importance of these two sources of coverage among all children who were uninsured in October 1992.

TABLE 13: DISTRIBUTION OF UNINSURED CHILDREN ELIGIBLE FOR MEDICAID IN OCTOBER 1992 BY HEALTH INSURANCE COVERAGE IN EACH OF NEXT 23 MONTHS
  Health Insurance Coverage in Calendar Month
Reference Month
and Calendar Month
Uninsured Medicaid- Eligible Uninsured Not Medicaid- Eligible Medicaid Employer- Sponsored Coverage Other Coverage Not in
Population
 
Total
 
0. October 1992 100 0 0 0 0 0 100
1. November 1992 79.5 7 7.4 5.3 0.2 0.6 100
2. December 1992 64.7 15.3 12.3 6.9 0.2 0.6 100
3. January 1993 61.6 13.4 14.8 9.3 0.2 0.9 100
4. February 1993 51.2 15.6 21.2 10.1 1.1 0.9 100
5. March 1993 47.2 16.3 22.9 12 0.7 0.9 100
6. April 1993 44.1 16.8 22.4 13.4 1.5 1.8 100
7. May 1993 42 18.1 22.1 14.3 1.5 2 100
8. June 1993 38 20 24.3 13.9 1.4 2.4 100
9. July 1993 38.2 21.2 24.2 11.6 1.8 3 100
10. August 1993 41.5 17.9 21.9 13.8 1.9 3 100
11. September 1993 40.1 19.1 22.5 12.4 2.3 3.6 100
12. October 1993 37.7 20 23.1 13.3 2.1 3.8 100
13. November 1993 41.2 15.8 22.3 14.9 2.2 3.6 100
14. December 1993 35.9 21 22.1 14.7 2.2 4 100
15. January 1994 36.6 16.8 23.1 16.9 2.5 4 100
16. February 1994 38 16.3 21.9 17.2 2.4 4 100
17. March 1994 34.4 18.6 23.2 17.6 2.2 4 100
18. April 1994 35 16.2 26.6 16.3 1.9 4 100
19. May 1994 34.4 17.2 25.8 16.8 1.6 4 100
20. June 1994 30.9 19 26.7 17.5 1.7 4.2 100
21. July 1994 31.1 19.1 26.4 16.6 2 4.7 100
22. August 1994 28.6 18.6 28.1 18.3 1.5 4.9 100
23. September 1994 26.3 21.2 26.9 18.2 2.5 4.9 100

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: The number of children under 19 who were uninsured but eligible for Medicaid in October 1992 is 

To highlight the impact of Medicaid eligibility, Table 14 presents the distribution of children who were uninsured but not eligible for Medicaid in October 1992 by their health insurance coverage in each of the next 23 months. These children left uninsurance somewhat more slowly than those who were initially eligible for Medicaid. After one month, 86 percent remained uninsured and not eligible for Medicaid while another 6.5 percent had gained Medicaid eligibility but remained uninsured. About 3 percent had enrolled in Medicaid and another 3 percent had obtained employer- sponsored coverage. These figures imply a Medicaid participation rate of about one third for children who are uninsured and newly eligible for Medicaid. Within a few months the Medicaid eligible are evenly split between those participating and those not participating in Medicaid. Toward the end of the 23-month period, the Medicaid participants are between one-and-a-half and two times the number of Medicaid-eligible nonparticipants, implying a Medicaid participation rate in excess of 60 percent.

Within four months of October 1992, children who have left the uninsured population are twice as likely to be covered by an employer-sponsored plan as Medicaid. This relative importance of the two sources of coverage among children who started out as uninsured and not eligible for Medicaid persists through September 1994.

c. Medicaid Enrollment

Table 15 reports the distribution of children enrolled in Medicaid in October 1992 by their health insurance coverage in each of the next 23 months. Children without health insurance coverage are apportioned between those who are eligible for Medicaid and those who are not. As reported in Table 10, about 3 to 4 percent of the children enrolled in Medicaid at any point in time leave the program within the next month. Table 15 shows that this rate of departures from Medicaid persists over as many as four months before slowing down. By the fourth month, 14 percent of the original Medicaid population has exited, with 5.5 percent picking up employer-sponsored coverage, .2 percent acquiring other coverage and an equal number leaving the survey population. Altogether 8 percent have become uninsured, with nearly 5 percent remaining Medicaid eligible. This latter percentage changes only marginally thereafter. The fraction of all October 1992 Medicaid participants classified as uninsured but still eligible for Medicaid never rises above 6.0 percent whereas the proportion who are uninsured but not eligible for Medicaid continues to ascend, gradually, reaching 8 percent by September 1994.

TABLE 14: DISTRIBUTION OF UNINSURED CHILDREN WHO WERE NOT ELIGIBLE FOR MEDICAID IN OCTOBER 1992 BY HEALTH INSURANCE COVERAGE IN EACH OF NEXT 23 MONTH
  Health Insurance Coverage in Calendar Month
Reference Month
and Calendar Month
Uninsured Not Medicaid- Eligible Uninsured Medicaid- Eligible Medicaid Employer- Sponsored Coverage Other Coverage Not in Population Total
0. October 1992 100 0 0 0 0 0 100
1. November 1992 86.2 6.5 2.9 3.4 0.5 0.5 100
2. December 1992 77.8 6.9 5.2 8.4 1.1 0.5 100
3. January 1993 69.5 7.4 7.2 12.5 2.3 1 100
4. February 1993 64.3 9.1 7.8 15.3 2.4 1.2 100
5. March 1993 62.8 8.4 7.6 17.5 2.6 1.2 100
6. April 1993 63.8 7.8 7.4 17.5 2.2 1.3 100
7. May 1993 60.2 8.1 8.6 19.6 2.2 1.3 100
8. June 1993 58.3 7.8 8.9 21.2 2.3 1.5 100
9. July 1993 58.9 7.6 8.8 21 2 1.6 100
10. August 1993 57.4 6.6 9.4 22.6 2.3 1.6 100
11. September 1993 55.8 7.2 10.2 22.8 2.5 1.5 100
12. October 1993 52.7 7.8 11.1 23.6 3.4 1.5 100
13. November 1993 51.8 7.9 11.6 23.5 3.8 1.5 100
14. December 1993 52.8 7.1 11.6 23.5 3.5 1.5 100
15. January 1994 52.8 7.6 11.7 22.7 3.6 1.6 100
16. February 1994 52.4 8 11.1 23.7 3 1.7 100
17. March 1994 51.6 8.7 10.8 24.2 3.1 1.7 100
18. April 1994 49.2 8.7 11.9 24.6 3.9 1.7 100
19. May 1994 48.5 8.8 11.9 24.9 4.1 1.8 100
20. June 1994 47.6 6.8 12.9 26 4.8 1.9 100
21. July 1994 46.1 6.7 14.2 25.5 5.2 2.3 100
22. August 1994 46.9 7.3 13.3 25.9 4.3 2.4 100
23. September 1994 45.7 7.5 12.5 27.6 4.2 2.5 100

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: The number of children who were uninsured and not eligible for Medicaid in October 1992 is 6,630,000.

TABLE 15: DISTRIBUTION OF CHILDREN ENROLLED IN MEDICAID IN OCTOBER 1992 BY HEALTH INSURANCE COVERAGE IN EACH OF NEXT 23 MONTHS
  Health Insurance Coverage in Calendar Month
Reference Month
and Calendar Month
Medicaid Employer- Sponsored Coverage Other Coverage Uninsured Medicaid- Eligible Uninsured Not Medicaid- Eligible Not in Population Total
0. October 1992 100 0 0 0 0 0 100
1. November 1992 96.5 1.7 0 1 0.7 0.1 100
2. December 1992 92.7 3.3 0.1 2.3 1.4 0.2 100
3. January 1993 89.4 4.3 0.1 3.8 2.2 0.4 100
4. February 1993 86.1 5.5 0.2 4.9 3.1 0.2 100
5. March 1993 86 5.4 0.4 4.5 3.6 0.2 100
6. April 1993 85.2 5.3 0.4 5 4 0.1 100
7. May 1993 82.8 6.4 0.4 6 4.1 0.2 100
8. June 1993 82.7 6.7 0.8 5.7 3.8 0.3 100
9. July 1993 81.1 7.1 0.9 5.6 4.9 0.5 100
10. August 1993 80.4 7.2 0.9 5.4 5.3 0.7 100
11. September 1993 79.8 7.1 0.9 4.8 6.6 0.8 100
12. October 1993 78.4 7.8 0.6 5.2 7.1 0.8 100
13. November 1993 78.2 8.6 0.6 5.2 6.6 0.8 100
14. December 1993 77.7 9 1 4.2 7.3 0.8 100
15. January 1994 77 8.7 1 5.2 7.1 1 100
16. February 1994 76.8 8.4 1.2 5.6 7.1 1 100
17. March 1994 75.8 9.1 1.4 5.1 7.6 1 100
18. April 1994 76.1 8.7 1.2 5.8 7.1 1.1 100
19. May 1994 75.6 8.6 1.2 5.7 7.6 1.3 100
20. June 1994 75.8 9.3 1 5 7.6 1.3 100
21. July 1994 75.2 9.3 1 5.8 7.4 1.3 100
22. August 1994 74.3 9.3 1.1 5.5 8.2 1.7 100
23. September 1994 74.1 9.6 1 5.5 8.1 1.7 100

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: The number of children under 19 enrolled in Medicaid in October 1992 is 12,614,000.

After four months, the proportion of October 1992 Medicaid children who are covered by employer-sponsored insurance stands at 5.5 percent. This proportion rises only another four percentage points over the next 19 months, reaching 9.6 percent by September 1994. In sum, the picture presented by Table 15 is one of very slow change in the population of children who start FY93 enrolled in Medicaid. Two years later nearly three-quarters of these children are still enrolled in Medicaid, with only 11 percent having moved to some other type of coverage and nearly 14 percent having become uninsured. More than a third of the uninsured appear to have retained their Medicaid eligibility. While this is puzzling, it is important to keep in perspective the small size of this group. At 5.5 percent of those who were enrolled in Medicaid in October 1992 they number 692,000, which makes them barely more than one quarter of the 2.6 million children who were uninsured and Medicaid-eligible in September 1994.

E. Churning Versus Turnover

The results presented in the preceding section indicate that about 600,000 children, on average, enrolled in Medicaid each month of FY93 and FY94 while nearly 800,000 became uninsured. Comparable numbers of children left Medicaid or became insured, so that the net change in the Medicaid child caseload or the population of uninsured children was small. Movements of children out of and into Medicaid or uninsurance include both “churning”--that is, exits entries by the same children over a relatively brief interval--and “turnover,” or exits and entries by different children.

Whether we classify a particular transition into Medicaid or uninsurance as an example of churning or not may depend on how we define the reference period for determining prior Medicaid enrollment or uninsurance. There’s no question that a child’s becoming uninsured after being covered for only six months should be included within the concept of churning.11 But what about a child who becomes uninsured after 13 months of coverage--or 24 months? Whatever we take churning to mean conceptually, moreover, measuring churning with available data may be quite another matter.

With data from the SIPP one could evaluate each transition into Medicaid or uninsurance within a period and determine whether and how long a child may have last been enrolled in Medicaid or uninsured. For transitions occurring in calendar year 1993, for example, one could look back as far as 12 months for children making transitions into Medicaid or uninsurance in January 1993 and as far back as 23 months for children making transitions in December 1993. For transitions occurring in calendar year 1994, one could look back 24 months for children making transitions in January 1994 and up to 35 months for children making transitions later in the year. Maintaining a consistent operational definition of churning would require that we look back no farther for children making transitions late in the period than we are able to do for children making transitions early in the period. This would imply that for transitions occurring in 1993 we look back at most 12 months and for transitions occurring in 1994 we look back at most 24 months.

Here we take a somewhat different approach to measuring the relative amount of churning and turnover in the transitions that we observe into Medicaid enrollment and uninsurance. We do so in order to pull together the findings that we presented earlier and in Technical Appendix A. From the data presented in Tables 9 and 11, we can count the number of transitions of children into Medicaid and uninsurance over a 23 month period, from November 1992 through September 1994. We also know the number of children enrolled in Medicaid or uninsured at the beginning of the period, October 1992 (or any month along the way). Finally, from data presented elsewhere, we know how many children under 19 were ever enrolled in Medicaid and how many were every uninsured in each of FY93 and FY94 and over the full two-year period. Subtracting enrollment or uninsurance at the beginning of a period from the number of children who were ever enrolled or ever uninsured during the period gives us a measure of aggregate turnover. Specifically, it tells us how many additional children entered Medicaid or uninsurance over the period, which we can express as a percentage of the initial enrollment. Subtracting the aggregate turnover from the total number of transitions into Medicaid or uninsurance over the period gives us a measure of excess transitions, which represent churning. In effect, then, we are counting each transition as an instance of either turnover or churning on the basis of whether the individual making the transition was in that state at the beginning of the period. This is not a time-based definition as we discussed above, but it provides a measure that is not only very intuitive but indicative of how much change occurred in the Medicaid or uninsured population over the period.

To obtain the population uninsured at the beginning of a period requires a small calculation.12 Children uninsured in the first month of a period include children who were uninsured before the period began and children for whom this is their first full month of uninsurance.13 Children ever uninsured in the period include those with one or more months of uninsurance. To count how many children became uninsured during the year we must compare the number ever uninsured during the year to those uninsured at the beginning of the year. If we equate all children who were uninsured in the first month--that is, the point-in-time estimate for that month--with the number uninsured at the beginning of the year, then our estimate of children who became uninsured during the year does not include those who became uninsured in the first month. To count children who became uninsured in the first month as well as the remaining 12, we need to redefine the population at the beginning of the year as children uninsured in the first month minus those who began spells of uninsurance in that month.14

Our estimates of the components of change in the population of uninsured children are presented in Table 16. For FY93 we start with about 9.5 million uninsured children in October 1992. From this we subtract the estimated 551,000 children who became uninsured in that month to obtain the number of children who were uninsured at the beginning of the year: 8.9 million.15 Altogether 16.1 million children were ever uninsured during FY93. Subtracting the initial population from this number yields our estimate of the number of children who became uninsured in FY93: approximately 7.2 million. This estimate of the turnover in the uninsured population represents 80 percent of the number who were uninsured at the start of the period and 44 percent of the number who were ever uninsured during the period. We obtain somewhat higher estimates of turnover in FY94, with the children who were added to the uninsured population in that year representing 90 percent of the number uninsured at the beginning of the year and 47 percent of those who were ever uninsured during the year. For the two-year period FY93 and FY94 we find that 12.1 million children became uninsured over this time or 136 percent of the number who were uninsured at the beginning of FY93 and 58 percent of the number who were ever uninsured during the two years. That the 12.1 million new uninsured is not 15 million (the sum of the FY93 and FY94 additions) implies that about 3 million of the 7.6 million children who were added to the uninsured after the beginning of FY94 were uninsured at some point during FY93. These children will figure into our estimate of churning over the two-year period.

TABLE 16: ESTIMATES OF CHANGE IN THE POPULATION OF UNINSURED CHILDREN, FY93 AND FY94
Description of Estimate FY93 FY94 FY93 and FY94
1. Uninsured in first month of period 9,489,000 9,080,000 9,489,000
2. Estimated new uninsured that month 551,000 697,000 551,000
3. Number uninsured at start of period [(1) - (2)] 8,938,000 8,383,000 8,938,000
4. Number ever uninsured in period 1,6089,000 1,5936,000 21,074,000
5. Children becoming uninsured during period [(4) - (3)] 7,151,000 7,553,000 12,136,000
    a. Percent of number uninsured at start of period [(5)/(3)] 80.0% 90.1% 135.8%
    b. Percent of number ever uninsured during period [(5)/(4)] 44.4% 47.4% 57.6%
6. New spells started during period 9,281,000 9,463,000 18,744,000
7. Excess of new spells over new uninsured during period [(6) - (5)] 2,130,000 1,910,000 6,608,000
    a. Percent of new spells [(7)/(6)] 23.0% 20.2% 35.3%
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Nearly 9.3 million new spells of uninsurance were started in FY93 and nearly 9.5 million in FY94, for a two-year total of 18.7 million. In FY93 the number of new spells exceed by 2.1 million the number of children who were added to the uninsured in that year. These excess spells were started by children with one or more other spells during the year--that is, spells that were already in progress at the beginning of the year or spells that began during the year. These excess spells represent 23 percent of all the new spells that started during the year. Similarly, for FY94 we estimate that 1.9 million of the 9.5 million spells of uninsurance that started during the year were associated with children who had one or more other spells of uninsurance during the year. For the two-year period, the number of excess spells is more than just the sum of the numbers of such spells in FY93 and FY94. The total of 6.6 million includes spells started by children with no other spells in FY94 but with spells in FY93 that ended before the start of FY94. Measured in this way, churning accounts for 35 percent of all the new spells that were started in FY93 and FY94. By implication, if we were to define churning in FY94 in terms of spells started by children who had other spells of uninsurance in either FY94 or FY93, rather than just FY94, the number of such spells would equal the difference between the 6.6 million two-year total and the 2.1 million spells started in FY93 by children who had one or more other spells of uninsurance during FY93. This difference of about 4.5 million represents 47 percent of the new spells started in FY94. In other words, 47 percent of the new spells of uninsurance started in FY94 represented second, third, or even higher order spells by children who had been uninsured once previously in either FY93 or FY94. If we were to look for earlier spells as far back as two years for all children who started new spells in FY94, rather than just those who started new spells at the end of FY94, we would obtain an even higher estimate of churning. On the other hand, if we were to limit our definition of churning to include only those new spells started within a year of the last spell, then the estimate of churning in FY94 would lie somewhere between 20 percent and 47 percent.16

Table 17 presents analogous estimates of the components of change in the population of children enrolled in Medicaid in FY93 and FY94. While the number of children enrolled in Medicaid at a point in time is about one-third higher than the number of children who are uninsured, both the number of new spells started during a year and the number of new enrollees that these spells represent are smaller than the corresponding numbers for the uninsured. In FY93, new enrollees represented 48 percent of the number enrolled at the beginning of the year and 32 percent of the number of children ever enrolled during the year. For FY94 these figures were 40 percent and 29 percent, respectively, while the new enrollees over the two-year period were 75 percent of the initial enrollees and 43 percent of the children ever enrolled during the period. Clearly, then, Medicaid enrollees are a more stable population than uninsured children, with fewer new children joining and those who do enroll staying longer. Except for FY93, however, the rate of churning appears comparable for Medicaid and uninsurance. While the new spells started by children with earlier spells of Medicaid enrollment during the year were only 14 percent of all new spells in FY93, they were 23 percent of all new spells in FY94--higher than the corresponding figure for the uninsured. Likewise, new spells started over the two-year period FY93 and FY94 by children with earlier spells of enrollment during that same period represented 33 percent of all new spells compared to 35 percent for uninsurance. As with uninsurance, if we were to define churning in FY94 in terms of spells started by children who had other spells of Medicaid enrollment in either FY93 or FY94, the number of such spells would equal the difference between the 4.4 million two-year total and the 980,000 spells started in FY93 by children who had one or more other spells of Medicaid enrollment during FY93. This difference of about 3.4 million represents 51 percent of the new spells started in FY94. That is, 51 percent of the new spells of Medicaid enrollment started in FY94 represented second, third, or even higher order spells by children who had been enrolled in Medicaid once previously in either FY93 or FY94.

TABLE 17: ESTIMATES OF CHANGE IN THE POPULATION OF CHILDREN ENROLLED IN MEDICAID, FY93 AND FY94
Description of Estimate FY93 FY94 FY93 and FY94
1. Enrolled in first month of period 12,614,000 13,236,000 12,614,000
2. Estimated new enrollees that month 583,000 567,000 583,000
3. Number enrolled at start of period [(1) - (2)] 12,031,000 12,669,000 12,031,000
4. Number ever enrolled in period 17,800,000 17,795,000 21,043,000
5. Children becoming enrolled during period [(4) - (3)] 5,770,000 5,126,000 9,012,000
    a. Percent of number enrolled at start of period [(5)/(3)] 48.0% 40.5% 74.9%
    b. Percent of number ever enrolled during period [(5)/(4)] 32.4% 28.8% 42.8%
6. New spells started during period 6,750,000 6,668,000 13,418,000
7. Excess of new spells over new enrollees during period [(6) - (5)] 980,000 1,542,000 4,406,000
    a. Percent of new spells [(7)/(6)] 14.5% 23.1% 32.8%
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Technical Appendix C: Characteristics of Children by Health Insurance Status

Introduction

Using data from the Survey of Income and Program Participation (SIPP), this report examines the characteristics of children under the age of 19 by their health insurance status. The report presents findings on broad types of health insurance coverage while focusing in particular on Medicaid eligibility and participation and on children who lack coverage altogether--the uninsured. We include the following demographic characteristics: age, race and Hispanic origin, family composition, metropolitan versus nonmetropolitan residence, and region. We also include three socioeconomic characteristics: family income relative to the federal poverty level, parents' employment, and parents' education.

Section A presents findings with respect to health insurance coverage at a point in time. Section B examines the characteristics of children by Medicaid eligibility and participation. Section C looks at the lack of insurance over time. Section D presents estimates of the annual incidence and duration of uninsurance, Medicaid eligibility, and Medicaid participation by single year of age, utilizing an approach that yields the cleanest possible age differentials. Finally, Section E summarizes our major findings and conclusions.

A. Insurance Coverage at a Point in Time

In this section we examine the characteristics of children by their insurance coverage at a point in time, utilizing a measure of coverage that differentiates among employer-sponsored insurance, Medicaid, other public or privately purchased insurance, and the lack of coverage. We begin by comparing the insurance coverage of children with what was reported for their parents. We then examine how insurance coverage varies by children's demographic and socioeconomic characteristics.

1. Parent's Insurance Coverage

Table 1 presents a distribution of parent's insurance coverage by the child's insurance coverage for each of three points in time: October 1992, September 1993, and September 1994 (representing the beginning, middle, and end of the two-year period FY93 through FY94). "Parent" here refers to the individual identified in the SIPP file as the child's parent. Thus the parental coverage reported in Table 1 refers to only one parent. There are characteristics reported in later tables for which we take account of both parents, if present, where the second parent is identified as the spouse of the first parent. For health insurance coverage, however, we felt that we could not easily combine the four-category reports of two parents to produce a summary measure of both parents' coverage.

Altogether 18 to nearly 21 percent of uninsured children have a parent who reports some insurance coverage, and 7 to 8 percent of uninsured children have a parent who reports being covered by Medicaid. Another 4 percent of children appear to have no parent in the household, so their parents' insurance coverage must be regarded as unknown--and perhaps irrelevant.

We suspect that much if not most of the parent-child discrepancy in reported Medicaid coverage is due to measurement error--either in the misreporting of the child as not covered by Medicaid or the misreporting of the parent as covered. There are very few circumstances under which a parent would qualify for Medicaid while a child under 19 would not. If a parent is eligible for Medicaid by virtue of receiving Supplemental Security Income (SSI), then the children are not automatically eligible. It is likely, however, that many children who are in this situation would be eligible for Medicaid on the basis of their family income, given that the one parent receiving SSI and his or her SSI payments would not be counted in the Medicaid eligibility determination. Another situation that could result in a parent being covered by Medicaid but not the child is where the child qualifies for consideration as an adult. In this case the child could be ineligible for Medicaid despite having a parent and one or more siblings who qualify. We doubt that such cases as these could be numerous enough to account for the estimated 7 to 8 percent of uninsured children who have a parent covered by Medicaid, however.

TABLE 1: PERCENTAGE DISTRIBUTION OF PARENT'S INSURANCE COVERAGE BY CHILD'S INSURANCE COVERAGE
Period and Child's Insurance Coverage Grand Total Subtotal Parent Is Insured Parent Is Uninsured No
Parent in
Household
Employer – sponsored Medicaid Other
October 1992              
All Children 71,549,000 84.5 65.1 15.3 4.2 13 2.5
    Insured 62,060,000 94.7 73.4 16.5 4.7 3.1 2.2
        Employer-sponsored 46,439,000 97.4 95.7 1.1 0.6 1.4 1.2
        Medicaid 12,614,000 83.8 6.2 76.7 0.9 9.7 6.4
        Other 3,007,000 97.3 10.7 1.9 84.8 1.3 1.3
Uninsured 9,489,000 18.3 10.4 7.4 0.5 77.8 3.9
September 1993              
All Children 70,868,000 84.3 64.3 15.9 4.1 13.3 2.4
    Insured 61,597,000 94.1 72.2 17.2 4.7 3.9 2.1
        Employer-sponsored 45,396,000 97.3 95.1 1.3 0.9 1.7 1
        Medicaid 13,369,000 82.7 7.6 74.5 0.6 11.6 5.7
        Other 2,831,000 96.6 10 1.2 85.4 1.8 1.6
Uninsured 9,271,000 19.8 12 7.4 0.4 75.9 4.3
September 1994              
All Children 69,935,000 84.6 64.4 16.2 4.1 13.2 2.2
    Insured 61,024,000 93.9 72 17.3 4.6 4.1 2
        Employer-sponsored 44,858,000 97.5 95.6 1.2 0.6 1.6 0.9
        Medicaid 13,259,000 81.4 6.1 75.1 0.3 13 5.6
        Other 2,907,000 96.6 8.9 2.1 85.5 1.9 1.5
Uninsured 8,911,000 20.6 11.8 8.4 0.3 75.4 4
SOURCE: Survey of Income and Program Participation, 1992 Panel.

If most of the reported instances of uninsured children with Medicaid-covered parents are indeed erroneous, we suspect that the reported lack of coverage of the child is where the problem lies. The Medicaid enrollment of children is underreported in the SIPP by 13 to 15 percent (see Technical Appendix D). Situations where Medicaid coverage is reported for a parent but not reported for all children may account for some of the underreporting among children. Certainly this possibility merits further investigation and, if borne out, could provide a basis for making a partial correction to the reported Medicaid coverage of children in the SIPP.

The 10 to 12 percent of uninsured children whose parents are reportedly covered by employer-sponsored insurance raise important questions as well. Could some of this, too, be due to error in the reporting of either the parent's coverage or the child's lack of coverage? If the parent is truly covered and the child is not, did the parent have no access to dependent coverage, or did the parent find it too expensive or otherwise choose not to obtain such coverage? These are not questions that SIPP data can answer, but they are questions that must be posed to other datasets and new data collection efforts if we are to understand the prevalence of uninsurance among children.

The 10 to 13 percent of Medicaid-covered children with uninsured parents is consistent with the expansion of Medicaid coverage aimed exclusively at children. That this percentage grows over time is also consistent with the newness of these expansions--participation in new programs tends to start slowly--and the fact that they have continued to make more and more children eligible over time.1

Finally, the consistent 6 percent of children with no parent in the household among children who are covered by Medicaid is noteworthy because it is three times as high as the rate at which all children under 19 appear to have no parents in the household. The rate among uninsured children is 4 percent or midway between the all-child rate and the Medicaid rate. This differential frequency of no-parent households among these different groups of children may reflect a causal impact of parental absence on both Medicaid participation and uninsurance. It is plausible, certainly, that the absence of parents from a child's household affects the likelihood that the child will qualify for Medicaid and, also, the likelihood that no private or other public insurance coverage will be provided.2 At the same time, however, the absence of a parent from the survey household may reduce the likelihood that all sources of insurance coverage for a child are reported. In other words, the apparent relationship between a child's uninsurance and the parents' absence from the child's household may be due in part to the child's insurance coverage being underreported when neither parent is present in the household.3 This might seem to be contradicted by the Medicaid finding, where children without parents are more likely than other children to be covered by Medicaid, but the mechanisms may be different.

2. Demographic Characteristics

In this section we examine differentials in children's health insurance coverage by five demographic characteristics: age, race and ethnicity (specifically, Hispanic origin), family composition, metropolitan residence, and region. The findings describe children in September 1994. Before presenting our findings we describe our measurement of the five characteristics and show how the population of children is distributed with respect to each of these characteristics.

a. Distribution of the Population

Table 2 presents the distribution of all children by the demographic characteristics that will be used to examine differentials in insurance coverage.

We divide children 0 through 18 into six age groups, using breaks that are relevant to Medicaid eligibility and, therefore, insurance coverage as well: infant, 1 to 5, 6 to 8, 9 to 10, 11 to 14, and 15 to 18. We separate the 9 to 10 group because these children phased into coverage under the poverty-related criteria over the two fiscal years that we examine. Specifically, none of the children aged 9 to 10 at the end of FY92 qualified for poverty-related coverage under the mandatory federal guidelines whereas all of these children met the age limit for poverty-related coverage by the end of FY94. Children 1 to 5 and 11 to 15 each account for 27 percent of the children under 19 represented by the SIPP panel in September 1994. Infants account for 4 percent while the other age groups include between 10 and 16 percent of all children.

We divide children into four categories on the basis of race and ethnicity. White non-Hispanic children account for 68 percent of the total. Black non-Hispanic children are 15 percent of the total while Hispanic children represent about 13 percent, and children of other racial groups--Asian and Pacific Islander and American Indian, Eskimo, and Aleut account for the remaining 4 to 5 percent.

TABLE 2: DISTRIBUTION OF CHILDREN UNDER 19 BY DEMOGRAPHIC CHARACTERISTICS: SEPTEMBER 1994
Demographic Characteristic Number Percent
All Children 69,935,000 100
Age of Child   100
    Infant (0) 2,943,000 4.2
    1 to 5 18,881,000 27
    6 to 8 11,423,000 16.3
    9 to 10 7,356,000 10.5
    11 to 15 18,770,000 26.8
    16 to 18 10,562,000 15.1
Race/Ethnicity of Child   100
    White Non-Hispanic 47,473,000 67.9
    Black Non-Hispanic 10,425,000 14.9
    Hispanic 8,907,000 12.7
    Other 3,130,000 4.5
Family Composition   100
    Both Parents Present 50,242,000 71.8
    Mother Only Present 16,607,000 23.7
    Father Only Present 1,521,000 2.2
    No Parent Present 1,565,000 2.2
Metropolitan Residence   100
    Metro 51,560,000 73.7
    Non-Metro 17,542,000 25.1
    Not Applicable 833,000 1.2
Region   100
    New England 2,990,000 4.3
    Middle Atlantic 9,968,000 14.3
    East North Central 12,543,000 17.9
    West North Central 6,081,000 8.7
    South Atlantic 11,390,000 16.3
    East South Central 4,782,000 6.8
    West South Central 7,679,000 11
    Mountain 3,358,000 4.8
    Pacific 11,143,000 15.9
SOURCE: Survey of Income and Program Participation, 1992 Panel.

We define family composition in terms of which parents are present in the household. Children living with both parents account for 72 percent of all children while children with only their mothers present represent 24 percent. The remaining four percent are divided evenly between children living with their fathers and children with no parent present. The latter will include older teens living on their own or with friends as well as younger children being raised by relatives or others who are not their natural or adoptive parents.

The SIPP files do not identify rural residents explicitly, so we have used metropolitan area residence to differentiate between a group including primarily urban children and a group including primarily rural children. The metropolitan category includes 74 percent of children while the non-metropolitan category includes 25 percent. The SIPP files assign 1.2 percent of children to a "not applicable" category, which is a Census Bureau designation that is not defined.

We used the census region classification to break down children by region of the country. The nine census regions and the states they include are:

New England: Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and Connecticut

Middle Atlantic: New York, New Jersey, and Pennsylvania

East North Central: Ohio, Indiana, Illinois, Michigan, and Wisconsin

West North Central: Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas

South Atlantic: Delaware, Maryland, District of Columbia, Virginia, West Virginia, North Carolina, South Carolina, Georgia, and Florida

East South Central: Kentucky, Tennessee, Alabama, and Mississippi

West South Central: Arkansas, Louisiana, Oklahoma, and Texas

Mountain: Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, and Nevada

Pacific: Washington, Oregon, California, Alaska, and Hawaii

While the SIPP file does not individually identify nine of the 50 states, leaving us to randomly assign children in each of three groups to individual states, only one of these nine states is not grouped with others from the same region.

Four of the nine regions include between 14 and 18 percent of the child population, with the East North Central region being the largest. Two of the regions--New England and the Mountain region--include less than 5 percent of children under 19. The remaining three account for 7 to 11 percent of all children.

b. Distribution of Insurance Coverage

Table 3 presents for each of the five demographic characteristics the distribution of insurance coverage by subgroup.

The inverse relationship between age and Medicaid coverage and the direct relationship between age and uninsurance are striking, but it is important to keep in mind the generally antecedent role of employer-sponsored and other insurance. The older the child, the more likely that the child is covered by either employer-sponsored or other insurance (generally privately purchased or public insurance other than Medicaid). The 71 percent of children 16 to 18 who are covered by either of these sources contrasts with 64 percent of infants and 65 percent of children 1 to 5. Among those children who might otherwise be uninsured, Medicaid is much more likely to be utilized by younger children (who are more likely to be eligible) than older children. At the extremes, Medicaid covers nearly three-quarters of the infants who lack other coverage but less than half of the 16- to 18-year- olds who, similarly, have no other coverage. The percentage of all children covered by Medicaid in each age group varies from 12 percent among children 16 to 18 to a maximum of 26 percent among infants, with the proportion rising monotonically as age decreases. The frequency of uninsurance, conversely, rises from 9 percent among infants to 16 percent among children 16 to 18.

TABLE 3: DISTRIBUTION OF CHILDREN'S INSURANCE COVERAGE BY DEMOGRAPHIC CHARACTERISTICS: SEPTEMBER 1994
Demographic Characteristic Employer- Sponsored Medicaid Other Uninsured Total
All Children 64.1 19 4.2 12.7 100
Age of Child          
    Infant (0) 61.2 26.3 3.2 9.3 100
    1 to 5 60.6 25.4 3.4 10.5 100
    6 to 8 64.1 20.3 3.7 11.9 100
    9 to 10 63.3 18 4.9 13.8 100
    11 to 15 67.4 14.5 4.5 13.6 100
    16 to 18 66.1 12.4 5.2 16.3 100
Race/Ethnicity of Child          
    White Non-Hispanic 74.1 10.7 5.2 10 100
    Black Non-Hispanic 41.8 44.2 1.4 12.5 100
    Hispanic 38.8 33.6 1.4 26.3 100
    Other 59.2 18.3 5.6 16.8 100
Family Composition          
    Both Parents Present 74.9 8.5 5 11.5 100
    Mother Only Present 35.8 47.6 1.8 14.7 100
    Father Only Present 55 21.5 2.3 21.3 100
    No Parent Present 26.8 47.4 2.8 23 100
Metropolitan Residence          
    Metro 65.5 18.5 3.7 12.2 100
    Non-Metro 60.3 19.8 5.6 14.3 100
    Not Applicable 58 26.4 0.8 14.7 100
Region          
    New England 76.1 13.6 3.2 7.1 100
    Middle Atlantic 66.2 21.6 2.9 9.2 100
    East North Central 67.5 19.5 2.4 10.6 100
    West North Central 70.2 14 7.9 7.9 100
    South Atlantic 66.1 17.9 3.5 12.5 100
    East South Central 58.1 23.9 2.1 16 100
    West South Central 58.5 14.8 5.1 21.5 100
    Mountain 65.6 17.9 4.9 11.7 100
    Pacific 56 22.3 6.1 15.6 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

There are very strong differentials by race and ethnicity (Hispanic origin). Non-Hispanic whites have the highest rate of coverage by employer-sponsored plans at 74 percent and the lowest rates of Medicaid coverage (about 11 percent) and uninsurance (10 percent). Non-Hispanic blacks have the highest rate of Medicaid coverage at 44 percent or four times the rate among whites. Blacks also have the second lowest coverage by employer-sponsored plans at 42 percent, but the rate of uninsurance among blacks is only 2.5 percentage points above that of whites. Hispanics report the highest rate of uninsurance by far at 26 percent, with the lowest rate of coverage by employer-sponsored plans, but 34 percent report Medicaid coverage--a rate that is three times that of non-Hispanic whites. In all coverage groups but other insurance the other races category falls between non-Hispanic whites and Hispanics--generally closer to whites. While coverage by other than employer-sponsored plans or Medicaid is quite low at 4 percent nationally, non-Hispanic whites and other races report much higher coverage--between 5 and 6 percent--than do Hispanics and non-Hispanic blacks.

Differentials by family composition are as great as those by race and ethnicity. Children with both parents present have the highest rates of employer-sponsored and other coverage at 75 percent and 5.0 percent, respectively, and the lowest rates of Medicaid coverage and uninsurance: 8.5 percent and about 12 percent. At the opposite end, children with no parent present have the lowest rate of employer coverage at 27 percent--barely one-third the rate for children with both parents present--and the highest rate of uninsurance at 23 percent or double that of children with both parents present. Children with no parent present share the highest rate of Medicaid coverage with children in mother- only families, at 47 to 48 percent. Children in mother-only families have a much lower rate of uninsurance than children in father-only or no-parent families. Compared to children with no parent present, children in mother-only families have a higher rate of employer-sponsored coverage. Compared to children in father-only families, children in mother-only families are covered by Medicaid at a rate that more than compensates for their lower employer-sponsored coverage.

Differences by metropolitan residence are marginal at best. Children in non-metropolitan areas have a higher rate of uninsurance by two percentage points. Their lower rate of employer-sponsored coverage is partially offset by marginally higher coverage rates for Medicaid and other insurance. Children who could not be classified as metropolitan or not are too few in number for their coverage rates to provide meaningful information, but their rates of employer-sponsored coverage and uninsurance are strikingly similar to those of non-metropolitan children.

Finally, regional differences in rates of uninsurance are quite substantial. New England has the lowest rate at 7.1 percent while the West South Central region has the highest rate at 22 percent or three times the rate observed in New England. In general, rates of uninsurance are highest in the South and the West. Much of the latter, we suspect, can be attributed to the high proportion Hispanic in the Western states. The role of Medicaid in these regional differences in rates of uninsurance is ambiguous--no doubt because higher than average rates of Medicaid participation can reflect either or both the prevalence of poverty and broader eligibility criteria. Employer-sponsored coverage varies over a range of 20 percentage points, with New England having the highest rate of coverage and the Pacific states having the lowest, followed closely by the East and West South Central states.

c. Distribution of Demographic Characteristics within Coverage Groups

Table 4 is a recasting of Table 3 with column percentages instead of row percentages. For each demographic characteristic it shows the percentage distribution of demographic subgroups within each class of insurance coverage. Distributions of demographic subgroups within coverage groups are of interest insofar as they indicate that the members of a particular coverage group--such as the uninsured--have a markedly different demographic composition or profile than other coverage groups. Such distributions are of value in determining how much of a specific deficiency in coverage could be addressed by a targeted strategy focusing on certain subgroups. In general, sizable differentials in coverage by demographic or other characteristics imply sizable differences in the profiles of coverage groups--and vice versa. When coverage differentials are confined to differences between small subgroups and the balance of the population, however, the profiles of different coverage groups may appear very similar. The absence of large differences in the profiles of the various coverage groups does not imply that the differentials are unimportant, necessarily, but it does indicate that a narrowly targeted policy will not have much impact on overall coverage.

Age distributions differ little across the coverage groups except for Medicaid, where children under 6 account for 42 percent of children reporting Medicaid coverage but only 25 to 29 percent of each of the other three insurance classes. Clearly the age profile of Medicaid enrollees is different from that of children with other sources of coverage--or no coverage at all--but these other coverage groups differ little from each other in their age composition. Racial composition varies substantially across the coverage groups, however. Hispanic children represent 26 percent of the uninsured and 23 percent of Medicaid enrollees but only 8 percent of children covered by employer-sponsored plans. Non-Hispanic blacks account for 35 percent of Medicaid enrollees but less than 15 percent of the uninsured and 10 percent of children with employer-sponsored plans. Non-Hispanic white children dominate every group except Medicaid enrollees. They account for more than half of the uninsured and more than three-quarters of those with employer-sponsored coverage but only 38 percent of the Medicaid population.

TABLE 4: DISTRIBUTION OF CHILDREN'S DEMOGRAPHIC CHARACTERISTICS BY INSURANCE COVERAGE: SEPTEMBER 1994
Demographic Characteristic Employer- Sponsored Medicaid Other Uninsured
Age of Child 100 100 100 100
    Infant (0) 4 5.8 3.2 3.1
    1 to 5 25.5 36.2 22.3 22.3
    6 to 8 16.3 17.5 14.4 15.3
    9 to 10 10.4 10 12.3 11.4
    11 to 15 28.2 20.6 29 28.7
    16 to 18 15.6 9.9 18.8 19.3
Race/Ethnicity of Child 100 100 100 100
    White Non-Hispanic 78.4 38.3 84.8 53.2
    Black Non-Hispanic 9.7 34.8 5 14.6
    Hispanic 7.7 22.6 4.2 26.2
    Other 4.1 4.3 6 5.9
Family Composition 100 100 100 100
    Both Parents Present 83.9 32.3 87 64.9
    Mother Only Present 13.3 59.7 10.3 27.4
    Father Only Present 1.9 2.5 1.2 3.6
    No Parent Present 0.9 5.6 1.5 4
Metropolitan Residence 100 100 100 100
    Metro 75.3 72.1 66.2 70.5
    Non-Metro 23.6 26.3 33.6 28.1
    Not Applicable 1.1 1.7 0.2 1.4
Region 100 100 100 100
    New England 5.1 3.1 3.3 2.4
    Middle Atlantic 14.7 16.3 10 10.3
    East North Central 18.9 18.4 10.3 15
    West North Central 9.5 6.4 16.5 5.4
    South Atlantic 16.8 15.4 13.7 16
    East South Central 6.2 8.6 3.5 8.6
    West South Central 10 8.6 13.6 18.5
    Mountain 4.9 4.5 5.6 4.4
    Pacific 13.9 18.7 23.5 19.5
SOURCE: Survey of Income and Program Participation, 1992 Panel.

The impact of group size is evident for family composition as well. While children in two-parent families have the lowest incidence of uninsurance, they nevertheless account for 65 percent of all uninsured children. Children from mother-only families dominate Medicaid due to the eligibility rules for families. They account for 60 percent of Medicaid enrollees compared to only 13 percent of those with employer-sponsored coverage. The father-only and no-parent groups do not account for more than 6 percent of any coverage group.

The only group in which children from non-metropolitan areas account for noticeably more than their share of all children is other insurance, where they represent 34 percent of the total versus 24 to 28 percent for other coverage groups. Finally, while individual regions account for varying shares of the different coverage groups, the variation across coverage groups is generally modest. Perhaps the most notable exception occurs in the West South Central region, which accounts for less than 9 percent of Medicaid enrollees but more than 18 percent of the uninsured.

3. Socioeconomic Characteristics

Differentials in children's insurance coverage by key socioeconomic characteristics are as sizable as those we have reported for demographic characteristics. Here we examine differentials by family poverty level, parents' employment, and parents' education as observed in September 1994. We begin by describing our measurement of the three characteristics and showing how children under 19 are distributed by each of the three. Following this we present distributions of insurance coverage by levels of each of these characteristics and then report the percentage distribution of children by poverty level, parents' employment, and parents' education within each category of insurance coverage.

a. Distribution of the Population

To calculate the family poverty level for each child, we used the September 1994 family income and family size, as constructed by the Census Bureau, and compared the family income to 1/12 the annual poverty threshold for a family of the size reported in that month.4 By this measure, 9 percent of children under 19 were in families below 50 percent of the poverty line, and nearly 12 percent were in families between 50 and 100 percent of poverty, yielding a total poverty rate close to 21 percent (see Table 5).5 A slightly larger percentage of children were in families between 100 and 200 percent of poverty while a slightly smaller fraction were in families between 200 percent and 300 percent of poverty. Just over one-third or 35 percent of children were in families above 300 percent of poverty.

We measured parents' employment by first determining whether each parent present in the child's household in September 1994 was employed full time, part time, or not at all in that month, and then we classified children as having one or both parents employed full time, neither parent employed full time but one or both parents employed part time, and neither parent working. Children with no parent identified as present in the household were classified separately. By this measure, 78 percent of children had one or both parents employed full time, 5 percent had parents employed only part time, 14 percent had no parents working, and 2 percent had no parents present in the household.

TABLE 5: DISTRIBUTION OF CHILDREN UNDER 19 BY SOCIOECONOMIC CHARACTERISTICS: SEPTEMBER 1994
Socioeconomic Characteristic Number Percent
Poverty Level 69,935,000 100
    Less than 50% FPL 6,459,000 9.2
    50% to < 100% FPL 8,114,000 11.6
    100% to < 200% FPL 16,465,000 23.5
    200% to < 300% FPL 14,325,000 20.5
    300% FPL or Greater 24,573,000 35.1
Parents' Employment 69,935,000 100
    1 or More Full Time 54,894,000 78.5
    Part Time Only 3,642,000 5.2
    No Working Parent 9,834,000 14.1
    No Parent Present 1,565,000 2.2
Parents' Education 69,935,000 100
    No Parent Present 1,565,000 2.2
    6 Years or Less 1,602,000 2.3
    7 to 11 Years 6,780,000 9.7
    12 Years 23,323,000 33.3
    Attended College 16,620,000 23.8
    4-Year College Degree 10,383,000 14.8
    Graduate Work 9,662,000 13.8
SOURCE: Survey of Income and Program Participation, 1992 Panel.

To classify children by their parents' education, we assigned each parent present in the child's household in September 1994 to one of six educational levels, based on years of schooling completed and whether the parent obtained a college degree. We then classified each child by the higher of either parent's level of education when both were present or by the one parent's level of education when only one was present. Children with no parent present were assigned to a separate category. Only 2 percent of children had parents with six or fewer years of schooling, and 10 percent had parents with at most 7 to 11 years of schooling. The largest group of children, at 33 percent, had at least one parent with exactly 12 years of schooling (and the other, if present, with 12 years or less). The remaining 52.5 percent of children had parents who attended college but did not graduate (24 percent), completed a four-year degree (15 percent), or went on to graduate school (about 14 percent).

b. Distribution of Insurance Coverage

Table 6 provides estimates of differential insurance coverage by family poverty level, parents' employment, and parents' education.

As we would expect, both employer-sponsored coverage and Medicaid are strongly related to the child's poverty level, but in opposite directions. Only 11 percent of children in families below 50 percent of poverty and 21 percent of those between 50 and 100 percent of poverty have employer-sponsored coverage. More than half of those in families between 100 and 200 percent of poverty have such coverage, and this proportion rises to 90 percent among children in families above 300 percent of poverty. For Medicaid we find that 70 percent of children below 50 percent of poverty and 53 percent of those between 50 and 100 percent of poverty are reported to be covered. Medicaid coverage declines to 20 percent among those between 100 and 200 percent of poverty then plunges to 5 percent for those between 200 and 300 percent of poverty and drops further to below 2 percent among children at 300 percent of poverty or higher.

TABLE 6: DISTRIBUTION OF CHILDREN'S INSURANCE COVERAGE BY SOCIOECONOMIC CHARACTERISTICS: SEPTEMBER 1994
Socioeconomic Characteristic Employer- Sponsored Medicaid Other Uninsured Total
Poverty Level          
    Less than 50% FPL 10.8 70.2 2.8 16.2 100
    50% to < 100% FPL 21.2 53.1 2.4 23.3 100
    100% to < 200% FPL 54.4 19.6 5.1 20.9 100
    200% to < 300% FPL 80 5.1 3.8 11.1 100
    300% FPL or Greater 89.6 1.9 4.7 3.8 100
Parents' Employment          
    1 or More Full Time 77 7.3 4.6 11.1 100
    Part Time Only 34.7 37.2 5.8 22.3 100
    No Working Parent 9.3 72.9 1.3 16.5 100
    No Parent Present 26.8 47.4 2.8 23 100
Parents' Education          
    No Parent Present 26.8 47.4 2.8 23 100
    6 Years or Less 11.1 45.8 0 43.1 100
    7 to 11 Years 24.1 52.6 0.5 22.9 100
    12 Years 56.9 24.1 3.1 15.9 100
    Attended College 73.4 12.2 4.7 9.7 100
    4-Year College Degree 85.2 2.9 6 6 100
    Graduate Work 86.2 2.7 7.3 3.8 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Other insurance shows little relationship to poverty level. Children in families below poverty are somewhat less likely to have other insurance coverage, but there is no variation among children above poverty. Uninsurance is inversely related to poverty level but not linearly: children below 50 percent of poverty have a lower rate of uninsurance than children between 50 percent and 200 percent of poverty. Clearly, the Medicaid coverage that is available to children below 50 percent of poverty explains this group's comparatively low rate of uninsurance.

Parents' employment has a very strong relationship with the child's insurance status as well. Children with at least one parent working full time have the highest rate of employer-sponsored coverage at 77 percent, the lowest rate of Medicaid coverage at 7.3 percent, and the lowest rate of uninsurance at 11 percent. Employer-sponsored coverage plunges to 9 percent, and Medicaid coverage rises to 73 percent when there is no working parent in the household. Uninsurance rises as well but to a level between that of children with parents employed full time versus part time. Children with no parent present have rates of both employer-sponsored coverage and Medicaid coverage that are between those of children whose parents work part time and children whose parents do not work at all, but they have the highest rate of uninsurance, at 23 percent.

Parents' education is highly related to the incidence of every one of the four coverage groups. Employer-sponsored coverage increases from 11 percent to 86 percent from the lowest education level (six years or less) to the highest (some graduate work). Medicaid coverage declines from 46 percent to 3 percent between the lowest and highest levels of education although children with parents educated only 7 to 11 years have a higher Medicaid participation rate (53 percent) than those with less educated parents. Other insurance coverage grows from 0 percent to 7 percent between the lowest and highest levels of education while the rate of uninsurance declines from 43 percent to 4 percent. For each of these four coverage groups this is by far the broadest range of rates seen for any of the demographic or socioeconomic variables. Given the nonlinear relationship between family poverty level and a child's lack of insurance, we speculate that the strength of the relationship between parents' education and their children's lack of insurance may reflect more than just differential access to health insurance. Certainly this merits further investigation, including the examination of education differentials in a multivariate context, but if the relationship holds up there are clear policy implications for efforts to increase enrollment in Medicaid and the new state Children's Health Insurance Program (CHIP) initiatives.

c. Distribution of Socioeconomic Characteristics within Coverage Groups

Table 7, which is analogous to Table 4, gives the percentage distribution of children by poverty level, parents' employment, and parents' education within each class of insurance coverage.

Children in families between 100 and 200 percent of poverty account for 39 percent of the uninsured. The size of this portion of the uninsured is important because it includes the children addressed most explicitly by the CHIP legislation. Older children between 50 and 100 percent of poverty also constitute a prime target of CHIP. Across all ages this group accounts for 21 percent of the uninsured. Somewhat surprisingly, however, uninsured children in families above 300 percent of poverty, at 10 to 11 percent of all uninsured children, are nearly as numerous as uninsured children in families below 50 percent of poverty, who account for 12 percent of the total.

It is particularly clear from Table 7 how Medicaid complements employer-sponsored and other insurance coverage. Children below 100 percent of poverty account for only 5.3 percent of all children with employer-sponsored coverage and 13 percent of those with other coverage, but they account for 67 percent of the children covered by Medicaid. Children between 100 and 200 percent of poverty account for relatively similar shares of all children covered by employer-sponsored plans, Medicaid, or other insurance. Children above 200 percent of poverty represent 75 percent of the children covered by employer-sponsored plans and 58 percent of the children covered by other insurance but only 9 percent of the children covered by Medicaid.

TABLE 7: DISTRIBUTION OF CHILDREN'S SOCIOECONOMIC CHARACTERISTICS BY INSURANCE COVERAGE: SEPTEMBER 1994
Socioeconomic Characteristic Employer - Sponsored Medicaid Other Uninsured
Poverty Level 100 100 100 100
    Less than 50% FPL 1.5 34.2 6.2 11.8
    50% to < 100% FPL 3.8 32.5 6.7 21.2
    100% to < 200% FPL 20 24.3 28.6 38.7
    200% to < 300% FPL 25.6 5.5 18.5 17.8
    300% FPL or Greater 49.1 3.5 39.9 10.5
Parents' Employment 100 100 100 100
    1 or More Full Time 94.2 30.1 86.7 68.7
    Part Time Only 2.8 10.2 7.2 9.1
    No Working Parent 2 54.1 4.5 18.2
    No Parent Present 0.9 5.6 1.5 4
Parents' Education 100 100 100 100
    No Parent Present 0.9 5.6 1.5 4
    6 Years or Less 0.4 5.5 0 7.7
    7 to 11 Years 3.6 26.9 1.2 17.4
    12 Years 29.6 42.4 24.8 41.7
    Attended College 27.2 15.3 26.8 18.1
    4-Year College Degree 19.7 2.3 21.4 6.9
    Graduate Work 18.6 2 24.2 4.1
SOURCE: Survey of Income and Program Participation, 1992 Panel.

With respect to parents' employment what stands out is the 69 percent of uninsured children who have at least one parent employed full-time. At the same time, children with parents working part-time account for only 19 percent of the uninsured. Children with no working parents account for 18 percent of the uninsured but 54 percent of those enrolled in Medicaid.

Children whose more educated parent had exactly 12 years of schooling constitute one-third of all children under 19, but they represent 42 percent of children enrolled in Medicaid and 42 percent of the uninsured. Interestingly, the uninsured are distributed almost identically below and above this modal educational level whereas Medicaid enrollees come disproportionately from children whose parents had less than 12 years of education. This latter group contributes only 4 percent of children with employer-sponsored coverage and 1 percent of those with other coverage

B. Medicaid Eligibility and Participation

The number of uninsured children who appear to be eligible for Medicaid, according to survey estimates, has raised concerns about the adequacy of Medicaid outreach. In Appendix B we presented findings from the 1992 SIPP panel that suggest that: (1) most uninsured children who become eligible for Medicaid remain eligible for very short periods of time, (2) more than two-thirds of all spells of Medicaid-eligible uninsurance are preceded or followed by periods of uninsurance without Medicaid eligibility, and (3) about one-third of spells of Medicaid-eligible uninsurance are preceded and followed by spells of uninsurance without Medicaid eligibility. The often transitory nature of Medicaid eligibility during spells of uninsurance may help to explain why many uninsured children who appear to be eligible for Medicaid do not enroll. In addition, we found that more than 40 percent of spells of Medicaid-eligible uninsurance were preceded or followed by Medicaid enrollment, and perhaps one fifth of such spells were preceded and followed by Medicaid enrollment. In other words, a significant share of Medicaid-eligible uninsured children appear to have initiated their spells of uninsurance by leaving Medicaid or to have ended their spells of Medicaid-eligible uninsurance by enrolling in Medicaid. For the first group, the key policy question is not why these children have not enrolled in Medicaid but, rather, why they left Medicaid. For the second group the key policy question is not why have these children not enrolled in Medicaid but why have they not enrolled sooner. Thus our analysis of the dynamics of uninsurance, Medicaid eligibility, and Medicaid participation suggests that inferences about the inadequacy of Medicaid outreach from the observation that three or four million uninsured children appear to be eligible for Medicaid at any point in time are overly simplistic and may be only partially supported by a closer examination of this population--particularly before and after the period of Medicaid eligibility. In the first subsection below we examine the demographic and socioeconomic characteristics of uninsured children who appear to be eligible for Medicaid, and in the second subsection we look at Medicaid participation rates among all eligible children by their demographic and socioeconomic characteristics.

1. Medicaid Eligibility among Uninsured Children

Table 8 looks at Medicaid eligibility among uninsured children by demographic characteristics, and Table 9 does so by socioeconomic characteristics. Each table presents the number uninsured and the percent of these who are Medicaid-eligible for each demographic or socioeconomic group.

TABLE 8: MEDICAID ELIGIBILITY AMONG UNINSURED CHILDREN UNDER 19, BY DEMOGRAPHIC CHARACTERISTICS: SEPTEMBER 1994
Demographic Characteristic Number Uninsured Percent Medicaid- Eligible Share of Medicaid- Eligible Share of Medicaid- Ineligible
All Children 8,911,000 32.9 100 100
Age of Child     100 100
    Infant (0) 275,000 68.3 6.4 1.5
    1 to 5 1,986,000 47.5 32.2 17.4
    6 to 8 1,360,000 42 19.5 13.2
    9 to 10 1,014,000 43.3 15 9.6
    11 to 15 2,559,000 19 16.6 34.7
    16 to 18 1,719,000 17.8 10.4 23.6
Race/Ethnicity of Child     100 100
    White Non-Hispanic 4,741,000 27 43.7 57.9
    Black Non-Hispanic 1,305,000 44.8 19.9 12
    Hispanic 2,338,000 36.7 29.2 24.8
    Other 527,000 39.8 7.1 5.3
    Family Composition     100 100
    Both Parents Present 5,783,000 26.6 52.4 71
    Mother Only Present 2,444,000 50.3 41.9 20.3
    Father Only Present 324,000 32.5 3.6 3.7
    No Parent Present 360,000 16.9 2.1 5
Metropolitan Residence     100 100
    Metro 6,284,000 34.5 74 68.8
    Non-Metro 2,504,000 28.4 24.3 30
    Not Applicable 123,000 42.2 1.8 1.2
Region     100 100
    New England 211,000 48.2 3.5 1.8
    Middle Atlantic 920,000 33.2 10.4 10.3
    East North Central 1,334,000 38.4 17.5 13.7
    West North Central 482,000 38.1 6.3 5
    South Atlantic 1,422,000 30.4 14.8 16.6
    East South Central 763,000 35.7 9.3 8.2
    West South Central 1,652,000 27.3 15.3 20.1
    Mountain 393,000 34.4 4.6 4.3
    Pacific 1,735,000 31.1 18.4 20
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 9: MEDICAID ELIGIBILITY AMONG UNINSURED CHILDREN UNDER 19, BY SOCIOECONOMIC CHARACTERISTICS: SEPTEMBER 1994
Socioeconomic Characteristic Number Uninsured Percent Medicaid- Eligible Share of Medicaid- Eligible Share of Medicaid- Ineligible
Poverty Level     100 100
    Less than 50% FPL 1,049,000 78.3 28 3.8
    50% to < 100% FPL 1,889,000 54.8 35.3 14.3
    100% to < 200% FPL 3,447,000 28.2 33.1 41.4
    200% to < 300% FPL 1,590,000 4.7 2.6 25.3
    300% FPL or Greater 936,000 3.4 1.1 15.1
Parents' Employment     100 100
    1 or More Full Time 6,118,000 24.1 50.2 77.7
    Part Time Only 814,000 44.4 12.3 7.6
    No Working Parent 1,619,000 64.2 35.4 9.7
    No Parent Present 360,000 16.9 2.1 5
Parents' Education     100 100
    No Parent Present 360,000 16.9 2.1 5
    6 Years or Less 690,000 38.7 9.1 7.1
    7 to 11 Years 1,549,000 43.1 22.7 14.8
    12 Years 3,716,000 34.1 43.1 41
    Attended College 1,614,000 26.7 14.7 19.8
    4-Year College Degree 618,000 24.4 5.1 7.8
    Graduate Work 363,000 25 3.1 4.6
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Each table also indicates what share of the Medicaid-eligible uninsured and the Medicaid-ineligible uninsured each demographic or socioeconomic group represents.

Overall, 33 percent of the uninsured were estimated to be eligible for Medicaid in September 1994.6 The sharpest differentials in this percentage are by age of child. The eligibility rate is 68 percent among infants, then drops to between 42 percent and 48 percent among children age 1 to 10, then drops further to 18 or 19 percent among older children. Children 1 to 5 represent 32 percent of Medicaid-eligible uninsured children while children 6 to 10 combine to represent 34.5 percent of this population. Infants represent 6 percent of the Medicaid-eligible uninsured. Children 11 to 15 represent less than 17 percent of the Medicaid-eligible uninsured but 35 percent--the largest share--of children who are uninsured but not Medicaid-eligible. Similarly, children 16 to 18 account for 24 percent of the Medicaid-ineligible uninsured but only 10 percent of the Medicaid-eligible uninsured.

Racial and ethnic differences in the proportion of uninsured children who are eligible for Medicaid reflect differences in family composition and income level. Non-Hispanic blacks have the highest eligibility rate at 45 percent despite the fact that blacks also have the highest rate of Medicaid coverage by far (Table 3). Hispanic children and those of races other than white or black have eligibility rates between 37 and 40 percent while non-Hispanic whites have a 27 percent eligibility rate. Because of their overall numbers, non-Hispanic whites account for the largest shares of both the Medicaid-eligible and Medicaid-ineligible uninsured at 44 percent and 58 percent, respectively. Hispanic children account for the next largest shares of both groups: 29 percent of the Medicaid-eligible and 25 percent of the ineligible uninsured.

Nearly 51 percent of uninsured children in mother-only families are eligible for Medicaid compared to 33 percent in father-only families, 27 percent in two-parent families and 17 percent among children living without their parents. Again, because of size, two-parent families account for the largest shares of both the Medicaid-eligible and ineligible uninsured at 52 percent and 71 percent respectively. Children in mother-only families represent 42 percent of the Medicaid-eligible uninsured and 20 percent of the Medicaid-ineligible. Children in father-only or no parent families account for 5 percent or less of either group.

Uninsured children living in metropolitan areas are somewhat more likely to be eligible for Medicaid than those living in non-metropolitan areas (35 versus 28 percent), and they account for 74 percent of the Medicaid-eligible uninsured and 69 percent of the Medicaid-ineligible uninsured.

Except for New England, the regions vary by only 11 percentage points in the proportion of their uninsured children who are Medicaid-eligible. Children in the West South Central region have the lowest eligibility rate at 27 percent while uninsured children in the two North Central regions have a 38 percent Medicaid eligibility rate, which is 10 percentage points below that of New England. Consistent with this fairly limited variation in Medicaid eligibility, each region's share of all Medicaid-eligible uninsured children is fairly comparable to its share of Medicaid-ineligible uninsured children.

Medicaid eligibility among uninsured children is highly related to family income level, as we would expect. We see in Table 9 that uninsured children below 50 percent of poverty have a 78 percent Medicaid eligibility rate. Family composition, family resources, and low eligibility thresholds in some states account for why this eligibility rate is not even higher. The Medicaid eligibility rate among children between 50 and 100 percent of poverty is 55 percent while the rate among children between 100 and 200 percent of poverty is about half that level. Children in each of the three lowest poverty classes account for similar shares of all the Medicaid-eligible uninsured: between 28 and 35 percent. Children in families between 100 and 200 percent of poverty account for the largest share of uninsured children who are not eligible for Medicaid--41 percent--while children in families between 200 and 300 percent of poverty represent 25 percent of the Medicaid-ineligible uninsured.

Parents' employment is inversely related to uninsured children's Medicaid eligibility. Uninsured children with at least one parent working full time have a 24 percent Medicaid eligibility rate compared to 44 percent for children with parents working only part time. Children with no working parent have a 64 percent Medicaid eligibility rate, and they account for 35 percent of all Medicaid-eligible uninsured children. Children with at least one parent working full time, who represent more than two-thirds of all uninsured children, account for 50 percent of the Medicaid-eligible uninsured and 78 percent of the ineligible uninsured. Children with no parent present have the lowest Medicaid eligibility rate at 17 percent and account for only 2 percent of the Medicaid-eligible uninsured and 5 percent of the ineligible uninsured.

Medicaid eligibility varies unevenly by parents' education. Children whose parents have at most 7 to 11 years of schooling have the highest Medicaid eligibility rate at 43 percent while children whose parents are college graduates have the lowest eligibility rate at 24 to 25 percent. Children whose parents have exactly 12 years of schooling have a Medicaid eligibility rate matching the national average but because of their numbers account for 43 percent of all Medicaid-eligible uninsured children and 41 percent of Medicaid-ineligible uninsured children. Shares of both populations decline as parents' education rises or falls from this median level.

Interpretation of these very high rates of Medicaid eligibility that we see among subgroups of the uninsured must be tempered by some of our findings from the examination of eligibility over time, which suggest that spells of Medicaid-eligible uninsurance frequently end with children enrolling or re-enrolling in Medicaid or by losing their eligibility, and that spells of Medicaid-eligible uninsurance typically last just a few months. Furthermore, it is important to recognize that Medicaid enrollment of children may be underreported by more than 2 million children at a point in time (see Technical Appendix D), which is about the number of uninsured children that we estimate to be eligible for Medicaid. Now, not all of the children whose Medicaid enrollment is unreported will have been reported as uninsured, and our simulations underestimate the number of children who are eligible for Medicaid. Nevertheless, it is undoubtedly correct to infer that the Medicaid eligibility rate among uninsured children is overstated in Tables 8 and 9. What implications this has for the observed differentials is not clear, however. It is plausible that Medicaid underreporting is greatest for subgroups with the highest percentages of Medicaid eligibles among their uninsured. If so, the differentials themselves may be overstated. On the other hand, if the underreporting of Medicaid coverage is due in large part to a stigma that respondents associate with participation, then respondents who belong to subgroups with low participation could feel this stigma most strongly, with the result that Medicaid participation is more likely to be unreported in subgroups with low participation than in subgroups with high participation.

2. Medicaid Participation Rates

Tables 10 and 11 present Medicaid participation rates in September 1994 by demographic and socioeconomic characteristics among all eligible children and among children with no other

TABLE 10: MEDICAID PARTICIPATION RATES BY DEMOGRAPHIC CHARACTERISTICS: SEPTEMBER 1994
Demographic Characteristic All Eligible Children Children with No Other Insurance
All Children 64.7 79
Age of Child    
    Infant (0) 57.5 78.2
    1 to 5 64.8 81.9
    6 to 8 65.9 77.5
    9 to 10 57.9 71
    11 to 15 70.3 81
    16 to 18 64.9 75.8
Race/Ethnicity of Child    
    White Non-Hispanic 53.7 75
    Black Non-Hispanic 78.6 87.8
    Hispanic 67.1 74.1
    Other 59.4 71.1
Family Composition    
    Both Parents Present 46 65.2
    Mother Only Present 75.7 85.6
    Father Only Present 55 69
    No Parent Present 87.3 90.3
Metropolitan Residence    
    Metro 65.8 78.8
    Non-Metro 61.8 79.5
    Not Applicable 62.9 77.6
Region    
    New England 51.4 75.3
    Middle Atlantic 73.3 85.7
    East North Central 69.5 80.4
    West North Central 52 78.3
    South Atlantic 62.4 80.5
    East South Central 65.6 78
    West South Central 57.3 66.7
    Mountain 63.2 77.1
    Pacific 66.8 78.7
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 11: MEDICAID PARTICIPATION RATES BY SOCIOECONOMIC CHARACTERISTICS: SEPTEMBER 1994
Socioeconomic Characteristic All Eligible Children Children with No Other Insurance
Poverty Level    
    Less than 50% FPL 75.4 84.2
    50% to < 100% FPL 67.3 79.6
    100% to < 200% FPL 50.2 68.6
    200% to < 300% FPL 45 78.7
    300% FPL or Greater 52 86.5
Parents' Employment    
    1 or More Full Time 40.7 63.7
    Part Time Only 64.4 78.9
    No Working Parent 82.2 86.2
    No Parent Present 87.3 90.3
Parents' Education    
    No Parent Present 87.3 90.3
    6 Years or Less 65.7 69.6
    7 to 11 Years 75.7 82.7
    12 Years 65.7 78.9
    Attended College 55.3 78.3
    4-Year College Degree 27.4 56.7
    Graduate Work 31.3 62.6
SOURCE: Survey of Income and Program Participation, 1992 Panel.

insurance coverage. The first participation rate is defined as the number of Medicaid-enrolled children who were simulated to be eligible divided by the number of all children who were simulated to be eligible.7 The second participation rate is defined as the number of Medicaid-enrolled children who were simulated to be eligible divided by the number of children who were simulated to be eligible and had no other source of insurance. That is, the participation rate is calculated as the number of Medicaid-enrolled children who are simulated to be eligible divided by the sum of these same Medicaid-enrolled children and Medicaid-eligible uninsured children.

Calculated over all eligible children, the estimated Medicaid participation rate is about 65 percent. When we exclude from the denominator those eligible children who are covered by insurance other than Medicaid, the estimated participation rate rises by 14 percentage points to 79 percent.8 Neither rate includes an adjustment for the underreporting of Medicaid participation in the SIPP, and, as explained above, neither rate includes reported participants who were simulated to be ineligible. Even with these caveats, which imply that the rates may be understated, the participation rate among children who would otherwise be uninsured is quite high. While the participation rate among all eligible children is significantly lower, the fact that the difference between the two rates is attributable solely to children who were reported to have other coverage puts the overall participation rate in a new light. Certainly we should not assume that this other coverage was necessarily more comprehensive than Medicaid, and in many cases almost surely it was not. But for a significant number of children who were Medicaid-eligible but not participating the choice faced by their parents appears to have been between Medicaid and another source of coverage rather than between Medicaid and no insurance at all. Why parents in this situation may have elected not to enroll their children in Medicaid is a question whose answer may have implications for Medicaid and CHIP outreach as well as for strategies to minimize the potential crowd-out effects of CHIP.

Comparison of the two columns of Table 10 shows that in addition to raising the participation rate by 14 percentage points, on average, the effect of excluding children with other insurance from a Medicaid participation rate is to render the rates much more equal across demographic subgroups. This is most striking for race and ethnicity, where non-Hispanic whites lag behind Hispanics by 13 percentage points and behind non-Hispanic blacks by 25 percentage points when the participation rate is calculated among all Medicaid-eligible children. When children with other insurance are excluded from the participation rate, whites show effectively the same participation rate as Hispanics (75 percent versus 74 percent) and lag behind blacks by only 13 percentage points. Puzzling age differences are largely eliminated when the rate is calculated just among children who would otherwise be uninsured, and differentials by family composition are reduced as well. A small metropolitan/non-metropolitan difference is eliminated, and the surprisingly low participation rate in New England is increased by 24 percentage points with the alternative measure. While regional differences remain, only one region--the West South Central--is very far out of line with the rest with a participation rate of 69 percent or 19 percentage points below the highest participation rate. In fact, if these two regions are eliminated, the remaining seven regions fall within a 5 percentage point range.

The story is much the same for socioeconomic differentials (Table 11). When all eligible children are included in the calculation, the range of participation rates among poverty classes is 31 percentage points. This drops to 20 percentage points when children with other insurance are excluded, and the relationship between poverty and Medicaid participation becomes U-shaped, with the lowest participation rate found in the middle poverty group and the highest rates found at the ends. Parents' employment shows a similar equalizing of rates with the alternative measure of participation. There remains a monotonic, inverse relationship between the level of parents' employment and their children's participation rate in Medicaid, but the range of participation rates is reduced from 46 percentage points to 26 percentage points. Finally, differentials in children's participation rates by parents' education level are reduced as well, with a 20 point differential across the middle three levels being reduced to 4 percentage points. The children of college graduates (including those who did graduate work) remain much less likely to participate than the children of less educated parents, but instead of being 45 percentage points below the group with the highest participation rate they are less than 25 percentage points below this group. Nevertheless, we continue to see evidence suggesting that the education differential reflects more than just access to health insurance. The 70 percent participation rate for children whose parents have no more than an elementary school education contrasts with participation rates of 84 percent for children below 50 percent of poverty and 86 percent for children with no working parent. The comparatively low participation rate of children in this lowest parents' education category underscores the need for further research on the relationship between education and insurance coverage.

Why do the differentials in participation rates decline when we exclude children with other coverage, and what does this tell us about participation in Medicaid? Some compression of the differentials may have occurred simply because we reduced the variance of participation by removing a group of nonparticipants from the rates. The dominant reason for the reduction, however, is that in removing children with other sources of coverage we have removed one source of the original differentials. That is, differences in Medicaid participation rates exist in part because children in different demographic and socioeconomic groups are differentially likely to have other coverage. That we observe these reductions in differential participation rates also suggest that it is unlikely that most of this other coverage is misreported Medicaid coverage. If it were largely Medicaid, then participation rates would be boosted more uniformly--unless, of course, the misreporting itself occurred disproportionately among groups that were least likely to participate. This is not implausible--particularly if the perceived stigma attached to Medicaid contributed to the misreporting, as well it might. The Medicaid stigma might also contribute to parents' decisions to choose an alternative source of coverage over Medicaid if such coverage is available. Clearly, we have identified an area where further research would be beneficial, as there could be important policy implications in the reasons why parents may have and choose alternatives to Medicaid, as we noted previously.

C. Lack of Insurance over Time

From looking at differentials in health insurance coverage at a point in time we turn now to differentials in the lack of health insurance coverage over time. First we compare children who were uninsured at the end of a year with those who were ever uninsured during the year. Next we look at differentials in the duration of all new spells of uninsurance that began during FY93, followed by new spells of uninsurance coupled with Medicaid eligibility. We close by examining differentials in the duration of spells that were active at the end of FY93 and comparing these to differentials in the duration of new spells that began during the year.

1. Uninsured At a Point in Time Versus Ever in a Year

Table 12 compares demographic differentials in the number of children uninsured in September 1994 and the number ever uninsured during FY94, and Table 13 compares socioeconomic

differentials. In both tables we see that the percentages of children ever uninsured during the year are much higher than the percentages uninsured in September 1994, and this in itself is striking, but the differentials are not materially different except for race/ethnicity and region. Black non-Hispanic children are more likely than children in the "other" group of races to have been ever uninsured during the year but less likely to be uninsured in September 1994. With respect to region there is a very clear division between the Northeast and Midwest, on the one hand, and the South and West on the other in the percentage ever uninsured during FY94. In the Northeast and Midwest the proportions ever uninsured during the year vary from 12 to 17 percent while in the South and West these proportions range between 22 and 33 percent. We note as well that the very large differentials in the frequency of uninsurance by parents' education are even more impressive with respect to children ever uninsured during the year. The proportion of children ever uninsured in FY94 rises from 8 percent among children whose parents did some graduate work after college to 54 percent among children whose parents received less than 7 years of education

Consistent with the generally very limited change in differentials between children uninsured in September 1994 and children ever uninsured during FY94, shares of the two groups of uninsured show little difference. This holds true even where the differentials in the proportion ever uninsured during the year are clearly larger than the differentials in the proportion uninsured in September 1994. For example, while the difference between Hispanic children and non-Hispanic white children grows from 16 percentage points to 20 percentage points, the Hispanic share of children who were ever uninsured is actually 3 to 4 percentage points lower than the Hispanic share of children uninsured in September 1994. Because the white population is so much larger than the Hispanic population, a smaller percentage point change can nevertheless add more white children than Hispanic children to the number of uninsured. We see a similar phenomenon with respect to region. Despite larger percentage point increases in the proportion uninsured in the South and West, the share of uninsured children living in these regions is no higher for children ever uninsured in FY94 than for children uninsured in September 1994.

TABLE 12: CHILDREN UNINSURED IN SEPTEMBER 1994 AND EVER IN FY94, BY DEMOGRAPHIC CHARACTERISTICS
Demographic Characteristic Percent Uninsured Share of Uninsured
In Sept. 1994 Ever in FY94 In Sept. 1994 Ever in FY94
All Children 12.7 21.5 100 100
Age of Child     100 100
    Infant (0) 9.3 14.2 3.1 2.8
    1 to 5 10.5 19.9 22.3 25.1
    6 to 8 11.9 20.9 15.3 15.9
    9 to 10 13.8 21.4 11.4 10.5
    11 to 15 13.6 22.3 28.7 27.9
    16 to 18 16.3 25.5 19.3 17.9
Race/Ethnicity of Child     100 100
    White Non-Hispanic 10 17.8 53.2 56.3
    Black Non-Hispanic 12.5 23.7 14.6 16.5
    Hispanic 26.3 38.3 26.2 22.7
    Other 16.8 21.4 5.9 4.5
Family Composition     100 100
    Both Parents Present 11.5 19.5 64.9 65.2
    Mother Only Present 14.7 25.1 27.4 27.8
    Father Only Present 21.3 33.7 3.6 3.4
    No Parent Present 23 34.8 4 3.6
Metropolitan Residence     100 100
    Metro 12.2 20.7 70.5 71.2
    Non-Metro 14.3 23.3 28.1 27.2
    Not Applicable 14.7 27.9 1.4 1.5
Region     100 100
    New England 7.1 11.9 2.4 2.4
    Middle Atlantic 9.2 15.3 10.3 10.1
    East North Central 10.6 17.3 15 14.5
    West North Central 7.9 15.3 5.4 6.2
    South Atlantic 12.5 22 16 16.7
    East South Central 16 26.6 8.6 8.5
    West South Central 21.5 33.3 18.5 17
    Mountain 11.7 24.9 4.4 5.6
    Pacific 15.6 25.7 19.5 19.1
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 13: CHILDREN UNINSURED IN SEPTEMBER 1994 AND EVER IN FY94, BY SOCIOECONOMIC CHARACTERISTICS
Socioeconomic Characteristic Percent Uninsured Share of Uninsured
In Sept. 1994 Ever in FY94 In Sept. 1994 Ever in FY94
Poverty Level     100 100
    Less than 50% FPL 16.2 27.6 11.8 11.9
    50% to < 100% FPL 23.3 34.2 21.2 18.5
    100% to < 200% FPL 20.9 32.4 38.7 35.5
    200% to < 300% FPL 11.1 20.5 17.8 19.5
    300% FPL or Greater 3.8 8.9 10.5 14.6
Parents' Employment     100 100
    1 or More Full Time 11.1 19.1 68.7 69.8
    Part Time Only 22.3 35.6 9.1 8.6
    No Working Parent 16.5 27.3 18.2 17.9
    No Parent Present 23 34.8 4 3.6
Parents' Education     100 100
    No Parent Present 23 34.8 4 3.6
    6 Years or Less 43.1 53.6 7.7 5.7
    7 to 11 Years 22.9 33.6 17.4 15.2
    12 Years 15.9 27.2 41.7 42.2
    Attended College 9.7 18.4 18.1 20.3
    4-Year College Degree 6 11.3 6.9 7.8
    Graduate Work 3.8 8 4.1 5.2
SOURCE: Survey of Income and Program Participation, 1992 Panel.

2. Duration of New Spells of Uninsurance

We have demonstrated that there are very pronounced demographic and socioeconomic differentials in the likelihood that a child is uninsured at a point in time or over the course of a year. Do these differences extend to the features of children's spells of uninsurance? In particular, are there important differences in the duration of spells of uninsurance by demographic or socioeconomic characteristics? We examined differentials in the distribution of completed spells of uninsurance among spells that started in FY93. Our findings with respect to demographic characteristics are presented in Tables 14 and 15. Socioeconomic differentials are reported in Tables 16 and 17.

As we explained in Appendix B, the durations of completed spells of uninsurance as measured in the SIPP incorporate a significant type of measurement error that greatly overstates the reported frequency of spells lasting exact multiples of four months. Consequently, the distributions reported in the tables that follow should not be interpreted literally. In particular, the 54 percent of all new spells that are reported as having been completed in 1 to 4 months is probably overstated, as some of the many spells identified as being completed in exactly 4 months undoubtedly lasted 5, 6, or even 7 months. Similarly, the percentage of spells lasting 13 months or more is probably understated, given the likely overreporting of spell durations of exactly 12 months. Intermediate spells are understated at the low end and overstated at the high end, with an uncertain effect on their overall frequency.9

TABLE 14: COMPLETED DURATION OF NEW SPELLS OF UNINSURANCE IN FY93, BY DEMOGRAPHIC CHARACTERISTICS AT START OF SPELL
Demographic Characteristic Completed Duration in Months Total
1 to 4 5 to 8 9 to 12 13+
All Children 54 16.9 9 20.1 100
Age of Child          
    Infant (0) 58.9 15.5 8.2 17.4 100
    1 to 5 56.7 14.2 9.3 19.8 100
    6 to 8 50.2 18.7 7.8 23.3 100
    9 to 10 53.2 19 8.3 19.4 100
    11 to 15 55.1 16.3 10.4 18.2 100
    16 to 18 49.8 20.2 8.1 21.8 100
Race/Ethnicity of Child          
    White Non-Hispanic 57 15.3 7.9 19.7 100
    Black Non-Hispanic 52 21.5 11.6 14.9 100
    Hispanic 50.1 16 8.8 25.2 100
    Other 50.6 18.9 10.2 20.3 100
Family Composition          
    Both Parents Present 49.7 18.2 8.9 23.2 100
    Mother Only Present 62.2 15 9 13.8 100
    Father Only Present 49.3 14.7 17.3 18.8 100
    No Parent Present 60.3 15.6 2.6 21.5 100
Metropolitan Residence          
    Metro 54.5 16.9 8.7 19.9 100
    Non-Metro 54.7 16.4 8.8 20.1 100
    Not Applicable 22.5 23.8 25.6 28.1 100
Region          
    New England 74.1 4.1 5.2 16.7 100
    Middle Atlantic 53.2 18.5 8.5 19.9 100
    East North Central 64.9 11.4 6.5 17.3 100
    West North Central 55.4 18.1 11.2 15.3 100
    South Atlantic 51.4 23.1 10.4 15 100
    East South Central 43.9 15.7 9.2 31.3 100
    West South Central 58.8 13.6 6.8 20.8 100
    Mountain 61.8 16.7 13.2 8.4 100
    Pacific 45.2 19.1 10.1 25.5 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 15: DISTRIBUTION OF DEMOGRAPHIC CHARACTERISTICS BY COMPLETED DURATION OF NEW SPELLS OF UNINSURANCE, FY93
Demographic Characteristic Completed Duration in Months
1 to 4 5 to 8 9 to 12 13+
Age of Child 100 100 100 100
Infant (0) 6.9 5.8 5.8 5.5
    1 to 5 29.8 23.7 29.4 28.1
    6 to 8 13.5 16.1 12.7 16.8
    9 to 10 9.6 11 9 9.4
    11 to 15 24.9 23.5 28.1 22.1
    16 to 18 15.4 19.9 15 18.1
Race/Ethnicity of Child 100 100 100 100
    White Non-Hispanic 53.3 45.7 44.5 49.7
    Black Non-Hispanic 20.4 26.9 27.3 15.8
    Hispanic 22.9 23.4 24.2 31
    Other 3.3 4 4 3.6
Family Composition 100 100 100 100
    Both Parents Present 55.1 64.4 59.4 69.3
    Mother Only Present 34.8 26.8 30.3 20.8
    Father Only Present 4.2 4 8.8 4.3
    No Parent Present 5.9 4.9 1.5 5.6
Metropolitan Residence 100 100 100 100
    Metro 75.1 74.4 71.9 73.7
    Non-Metro 24.3 23.3 23.5 24
    Not Applicable 0.7 2.3 4.6 2.3
Region 100 100 100 100
    New England 4 0.7 1.7 2.4
    Middle Atlantic 11 12.2 10.5 11
    East North Central 17.4 9.7 10.4 12.4
    West North Central 6.4 6.7 7.8 4.8
    South Atlantic 15.6 22.4 19.1 12.2
    East South Central 6.6 7.5 8.3 12.6
    West South Central 15.5 11.4 10.8 14.8
    Mountain 5.2 4.5 6.7 1.9
    Pacific 18.4 24.8 24.7 27.8
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 16: COMPLETED DURATION OF NEW SPELLS OF UNINSURANCE IN FY93, BY SOCIOECONOMIC CHARACTERISTICS AT START OF SPELL
Socioeconomic Characteristic Completed Duration in Months  
1 to 4 5 to 8 9 to 12 13+ Total
Poverty Level          
    Less than 50% FPL 62.8 15.5 8.1 13.7 100
    50% to < 100% FPL 55.3 17 8.6 19.1 100
    100% to < 200% FPL 50.8 17.8 8.6 22.8 100
    200% to < 300% FPL 51.2 13.6 11.4 23.8 100
    300% FPL or Greater 53 19.7 9.5 17.8 100
Parents' Employment          
    1 or More Full Time 53.5 17.2 8.5 20.8 100
    Part Time Only 49.8 20.2 13.1 16.8 100
    No Working Parent 56.3 14.5 10.5 18.7 100
    No Parent Present 60.3 15.6 2.6 21.5 100
Parents' Education          
    No Parent Present 60.3 15.6 2.6 21.5 100
    6 Years or Less 46 13.9 7 33.1 100
    7 to 11 Years 54.9 11.1 11.6 22.4 100
    12 Years 50.5 20.2 8.3 21 100
    Attended College 56.2 16 10.8 17 100
    4-Year College Degree 63.9 13.1 8.8 14.2 100
    Graduate Work 54.9 26 6.6 12.5 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.
TABLE 17: DISTRIBUTION OF SOCIOECONOMIC CHARACTERISTICS BY COMPLETED DURATION OF NEW SPELLS OF UNINSURANCE, FY93
Socioeconomic Characteristic Completed Duration in Months
1 to 4 5 to 8 9 to 12 13+
Poverty Level 100 100 100 100
    Less than 50% FPL 17.8 14 13.8 10.4
    50% to < 100% FPL 24.5 24 22.8 22.7
    100% to < 200% FPL 34.3 38.3 34.9 41.4
    200% to < 300% FPL 12.2 10.3 16.3 15.3
    300% FPL or Greater 11.3 13.4 12.2 10.2
Parents' Employment 100 100 100 100
    1 or More Full Time 65.8 67.6 62.5 68.9
    Part Time Only 9.1 11.8 14.5 8.3
    No Working Parent 19.2 15.8 21.5 17.1
    No Parent Present 5.9 4.9 1.5 5.6
Parents' Education 100 100 100 100
    No Parent Present 5.9 4.9 1.5 5.6
    6 Years or Less 5 4.8 4.6 9.7
    7 to 11 Years 18 11.6 22.8 19.7
    12 Years 33.8 43 33.4 37.7
    Attended College 22.6 20.5 26.1 18.3
    4-Year College Degree 8.4 5.5 6.9 5
    Graduate Work 6.4 9.7 4.7 3.9
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Despite this measurement error in the reported durations of spells, differentials in the relative frequency of spells of different length may carry important information about differences in the experience of uninsurance among children across demographic and socioeconomic subgroups. We recognize, though, that measurement error in the reporting of durations may vary by these same characteristics and either weaken or distort the observed relationships.

The results reported in Table 14 suggest that, generally, demographic differentials in the reported duration of children's spells of uninsurance are weak at best. There are no meaningful differences by age. With respect to race and ethnicity we find that Hispanic children have the fewest spells of 1 to 8 months in length and the most spells of 13 months or longer, but the differences between Hispanic children and other children are modest. Turning to family composition, we find that children in mother-only and no parent families have the highest incidence of spells reported to have lasted 4 months or fewer while children in mother-only families also have the lowest incidence of spells exceeding 12 months. There are no differences by metropolitan residence but there are differences by region. Spells tend to be shortest in New England and longest in the East South Central and Pacific regions. The size of the Hispanic populations in these latter areas may contribute to the longer duration of spells, given that Hispanic children appear to have the highest frequency of long spells. Finally, the East North Central and mountain regions join New England in having relatively short durations, but beyond this the regional differences are inconsistent.

With relatively weak differentials in the duration of uninsurance spells by demographic characteristics, we would not expect the distribution of demographic subgroups of uninsured children to vary much by spell duration. Table 15 shows the distribution of demographic subgroups of uninsured children within each grouping of spells by duration. Age composition shows no meaningful variation by spell duration. Race and ethnic composition show very limited variation. Non-Hispanic white children represent a somewhat larger share of 1 to 4 month spells than they do of longer durations while Hispanic children represent a larger share of spells exceeding 12 months than they do of shorter spells. Non-Hispanic black children appear with relatively greater frequency among spells of 5 to 12 months than they do among shorter or longer spells. Children in mother-only families represent 35 percent of 1 to 4 month spells compared to only 21 percent of 13 month or longer spells. Children in two-parent families show the opposite tendency. They represent 69 percent of the longest spells but only 55 percent of the shortest spells. Metropolitan residence continues to show no consistent variation while there are small regional differences. Children from New England and the East North Central states appear with greater frequency among 1 to 4 month spells than among longer spells while children from the Pacific states account for only 18 percent of the 1 to 4 month spells but 28 percent of the 13 month and longer spells.

Like demographic characteristics, socioeconomic characteristics display a much weaker association with spell duration than with the incidence of uninsurance. In Table 16 we see that poverty has a weak, inverse relationship with spell duration. Children from the poorest families tend to experience a higher frequency of very short spells and a lower frequency of long spells than children from less poor or higher income families. We speculate that this may reflect the impact of Medicaid as a source of coverage that is more available to children in very poor families than it is to children in less poor families. The highest incidence of spells lasting 13 months or more occurs

among children between 100 and 300 percent of poverty. Long spells decline in frequency in the top income class. Spell length shows a very inconsistent relationship with parents' employment, with no category being associated unambiguously with long spells or short spells. For example, children with no parent present are most likely to have spells of 1 to 4 months and spells of 13 months or longer. To some extent this inconsistency may be due to the fact that parents' employment is itself a dynamic characteristic. Differentials in the duration of uninsurance may depend on when we measure parents' employment (before, during, or after a spell of uninsurance). In Table 16, as the title indicates, we measure characteristics at the start (that is, in the first month) of a spell, which gives us a uniform measure across spells and increases the likelihood that the characteristics we observe are associated with why the child is uninsured. If we measure socioeconomic characteristics at only one point in time, clearly this is he preferred point to do so. But this is not to say that there is nothing to be learned by examining differentials in duration by income and parents' education measured at another point in time.

The educational attainment of children's parents has a moderately strong, inverse relationship to the duration of spells of uninsurance. This is most evident in the proportion of children with durations of 13 months or longer, which declines monotonically from 33 percent to 12 percent as parents' education rises (excluding the children with no parent present). It shows up as well in the proportion of children with spell durations of 8 months or less, which rises--also monotonically--from 60 percent to 81 percent as parents' education increases.

Socioeconomic differentials are even less evident in the shares of children completing spells of different durations (Table 17). Children in families below 50 percent of poverty represent an increasingly smaller share of spells as duration rises, but there are no other consistent patterns by poverty level. This lack of consistency is true of differences in shares of spells represented by children classified by their parents' employment and parents' education. The fact that we see such weak differences in spell shares by parents' education despite the fairly strong differences in the distribution of spell length by parents' education appears to be due to the small numbers of uninsured children whose parents fall into the lowest and highest levels of education. Children whose parents completed less than 7 years of schooling account for twice as many of the longest spells as they do all shorter spells, but this represents only a five percentage point difference. Similarly, children whose parents completed some graduate school represent only half as many of the longest spells as they do spells of 8 months or less, but this is only a 4 percentage point difference.

3. Duration of New Spells of Medicaid-Eligible Uninsurance

As with all spells of uninsurance, the duration of spells of Medicaid-eligible uninsurance shows similarly weak relationships to the demographic and socioeconomic characteristics of uninsured children. Part of the explanation is the very high proportion of spells that are completed in 4 months or less: 75 percent (Table 18). This leaves little room for significant variation across population subgroups.

Only infants show a pattern that differs to any degree from that of other age groups. Infants are somewhat less likely to have spells of one 1 to 4 months than the other age groups--65 percent versus about 76 percent--but they are no more likely than other groups to have very long spells. Non-Hispanic white children have both the highest frequency of very short spells, at 81 percent, and the lowest frequency of very long spells, at just under 5 percent. Other differences by race and ethnicity are negligible. Children in two-parent or father-only families also have the highest frequency of very short spells, at 77 to 79 percent, and the lowest frequency of very long spells, at 5 percent. The corresponding percentages for children in mother-only families are 70 percent and 9 percent, respectively. Children from no parent families compare at 70 percent and 10 percent.

TABLE 18: COMPLETED DURATION OF NEW SPELLS OF MEDICAID-ELIGIBLE UNINSURANCE IN FY93, BY DEMOGRAPHIC CHARACTERISTICS AT START OF SPELL
Demographic Characteristic Completed Duration in Months Total
1 to 4 5 to 8 9 to 12 13+
All Children 74.8 12.5 6.4 6.2 100
Age of Child          
    Infant (0) 65.5 18.8 9.1 6.6 100
    1 to 5 74.7 12.3 5.7 7.3 100
    6 to 8 73.1 17.1 7.2 2.6 100
    9 to 10 76.8 9.7 7.2 6.3 100
    11 to 15 79.1 8.1 5.3 7.5 100
    16 to 18 74.9 12 7.7 5.4 100
Race/Ethnicity of Child          
    White Non-Hispanic 80.8 10.1 4.4 4.7 100
    Black Non-Hispanic 66.3 18.5 6.9 8.3 100
    Hispanic 70 12.1 10.6 7.2 100
    Other 69.1 18.8 3.3 8.8 100
Family Composition          
    Both Parents Present 77 11.2 7 4.9 100
    Mother Only Present 69.7 15.9 5.3 9 100
    Father Only Present 78.7 7.6 8.5 5.2 100
    No Parent Present 69.8 17.8 2.2 10.2 100
Metropolitan Residence          
    Metro 72.4 13 8 6.7 100
    Non-Metro 81.9 10.7 2.6 4.7 100
    Not Applicable 54 29.2 5.9 10.9 100
Region          
    New England 74.6 13.8 0 11.7 100
    Middle Atlantic 61.2 19.8 4.4 14.7 100
    East North Central 72 7.9 12 8.2 100
    West North Central 80.9 19.1 0 0 100
    South Atlantic 77.7 15.5 4.2 2.6 100
    East South Central 69.4 9.6 8.4 12.6 100
    West South Central 76.4 13.9 4.4 5.4 100
    Mountain 67.6 8.5 14.8 9.1 100
    Pacific 80 9 7.5 3.4 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

There is a 10 percentage point difference between metropolitan and non-metropolitan children in the proportion of spells completed in 1 to 4 months. The children to whom the metropolitan/non-metropolitan classification is not applicable differ from the other two groups in the proportion of spells that are completed in 1 to 4 versus 5 to 8 months and in the proportion running 13 months or longer, but the uncertainty about who this group represents makes it unclear what this is telling us.

Because the number of children experiencing spells of Medicaid-eligible uninsurance is much smaller than the number who experience any spells of uninsurance, sampling error makes a greater contribution to the differentials in these tables than the earlier tables, and this is nowhere more

evident than in the regional differences. The West North Central and Pacific regions have the highest percentages of very short spells, at 80 to 81 percent, and among the lowest frequencies of very short spells, at 0 to 3 percent. The South Atlantic and West South Central regions are similar. Unlike all of the earlier tables, the New England region shows nothing distinctive in its patterns. In general, the regional patterns in Table 18 deviate so much from what we have seen earlier that we should be cautious in attaching much credence to them.

Table 19 reports shares of spells within each duration group by demographic characteristics. In general, we see little variation across the columns. Non-Hispanic white children account for 56 percent of the shortest spells compared to 39 percent of the longest spells while non-Hispanic black children represent 26 percent of the longest spells compared to 17 percent of the shortest spells. Similarly, children in two-parent families account for 66 percent of the shortest durations compared to 51 percent of the longest durations while children from mother-only families account for 26 percent of the shortest durations but 40 percent of the longest. A comparison with Table 8 indicates that these findings are consistent with the relative frequency of Medicaid eligibility among uninsured children by race and ethnicity and family composition although they appear to run counter to the

TABLE 19: DISTRIBUTION OF DEMOGRAPHIC CHARACTERISTICS BY COMPLETED DURATION OF NEW SPELLS OF MEDICAID-ELIGIBLE UNINSURANCE, FY93
  Completed Duration in Months
Demographic Characteristic 1 to 4 5 to 8 9 to 12 13+
Age of Child 100 100 100 100
    Infant (0) 5.7 9.8 9.2 6.9
    1 to 5 38.8 38.2 34.3 45.8
    6 to 8 17.3 24.1 19.9 7.5
    9 to 10 10.1 7.6 11 9.9
    11 to 15 19.2 11.7 14.9 22.1
    16 to 18 9 8.6 10.7 7.8
Race/Ethnicity of Child 100 100 100 100
    White Non-Hispanic 55.5 41.3 35 39
    Black Non-Hispanic 16.8 27.9 20.4 25.5
    Hispanic 24.1 24.8 42.6 29.9
    Other 3.7 5.9 2 5.7
Family Composition 100 100 100 100
    Both Parents Present 66.5 57.6 70.1 50.7
    Mother Only Present 25.7 35.1 22.7 40.2
    Father Only Present 4.8 2.8 6.1 3.8
    No Parent Present 3 4.5 1.1 5.3
Metropolitan Residence 100 100 100 100
    Metro 68.2 72.9 87.3 76.2
    Non-Metro 30.8 24.1 11.5 21.5
    Not Applicable 0.9 3.1 1.2 2.3
Region 100 100 100 100
    New England 2.6 2.9 0 5
    Middle Atlantic 8 15.5 6.7 23.3
    East North Central 13.8 9 26.7 19
    West North Central 6.8 9.5 0 0
    South Atlantic 18.3 21.7 11.6 7.3
    East South Central 7 5.8 9.9 15.4
    West South Central 14.6 15.9 9.7 12.4
    Mountain 2.6 1.9 6.5 4.2
    Pacific 26.3 17.7 28.9 13.5
SOURCE: Survey of Income and Program Participation, 1992 Panel.

differentials that we find in Medicaid participation (see Table 3). Clearly, a high participation rate in Medicaid does not imply that children leave spells of Medicaid-eligible uninsurance relatively quickly.

Metropolitan area children account for a bigger share of long spells than short spells at 76 percent versus 68 percent while non-metropolitan area children account for 31 percent of the shortest spells and 22 percent of the longest. Regional differences again show the effects of particularly large sampling error. Every region displays fairly wide variation in its shares of spells of different durations, but these differences do not yield a clear pattern.

Differences in the duration of spells of Medicaid-eligible uninsurance by socioeconomic characteristics are much weaker than the differences we reported in the duration of all spells of uninsurance (compare Table 20 with Table 16). There are negligible differences across the first three poverty classes, where sample sizes are largest, and the two highest poverty classes are not substantially different, given their very small numbers.10 Children with working parents, whether full time or part time, tend to have a greater frequency of short spells than children with no working parents or no parents present and a somewhat lower frequency of very long spells. Differences in spell duration by parents' education are very inconsistent, suggesting that what differences we do see are strongly affected by sampling error. Children whose parents attended graduate school clearly have the shortest spells, with 82 percent being 1 to 4 months in length and only 1 percent exceeding 8 months. Children whose parents completed less than 7 years of schooling have the longest spells, with 60 percent being 1 to 4 months and 23 percent 9 months or longer. But children whose parents completed college (and went no farther) look more like children whose parents completed only 7 to 11 years of schooling than they resemble children whose parents attended but did not complete college or completed college and went on to graduate school.

TABLE 20: COMPLETED DURATION OF NEW SPELLS OF MEDICAID-ELIGIBLE UNINSURANCE IN FY93, BY SOCIOECONOMIC CHARACTERISTICS AT START OF SPELL
  Completed Duration in Months  
Socioeconomic Characteristic 1 to 4 5 to 8 9 to 12 13+ Total
Poverty Level          
    Less than 50% FPL 74.9 15.5 4.6 5.1 100
    50% to < 100% FPL 74.7 12.7 7.1 5.5 100
    100% to < 200% FPL 75.5 10.4 7.1 7 100
    200% to < 300% FPL 67.8 18.4 3 10.7 100
    300% FPL or Greater 75.8 3 11.3 9.9 100
Parents' Employment          
    1 or More Full Time 78.4 11.6 4.4 5.7 100
    Part Time Only 74.9 11.4 9.5 4.3 100
    No Working Parent 66.6 14.9 10.4 8.1 100
    No Parent Present 69.8 17.8 2.2 10.2 100
Parents' Education          
    No Parent Present 69.8 17.8 2.2 10.2 100
    6 Years or Less 59.9 16.9 19.4 3.8 100
    7 to 11 Years 72.8 10.7 6.7 9.8 100
    12 Years 74.7 13.7 6.1 5.6 100
    Attended College 81.8 11 2.5 4.8 100
    4-Year College Degree 72.9 5.6 14.6 6.9 100
    Graduate Work 81.9 16.8 1.2 0 100
SOURCE: Survey of Income and Program Participation, 1992 Panel.

There are somewhat greater differences in the shares of spells of different durations by socioeconomic characteristics (Table 21) than by demographic characteristics (Table 19). Children in families below 50 percent of poverty account for smaller shares of spells 9 months or longer than they do of spells completed in 8 months or less, but patterns at higher family income levels are less clear. Children in families above 100 percent of poverty represent somewhat larger shares of the spells lasting 9 months or longer than they do of spells ending in 8 months or less. Children without working parents account for larger shares of spells 9 months or longer than they do of shorter spells while children with at least one parent working full time show the reverse pattern, but even here sampling error is evident in the volatility of the shares. Finally, children whose parents completed 7 to 11 years of schooling account for 33 percent of the longest durations compared to no more than 22 percent of any class of shorter durations whereas children whose parents attended but did not complete college represent 14 percent of the longest durations compared to 20 percent of the shortest.

4. Duration of Active Spells of Uninsurance

Earlier in this section we examined demographic and socioeconomic differentials in the lengths of completed spells of uninsurance. Spells that are active at a point in time have a different distribution--both currently and, especially, when completed--than do spells that started during a year. Furthermore, because active spells are generally not complete (some will end in that month, but most will continue), the relationships beween their duration and various characteristics of the uninsured differ from what we would observe with completed spells. More specifically, the shorter the active duration, the greater the proportion of these spells that will end up with longer completed durations. In terms of their characteristics, children who are in short active spells will look much more like children with completed spells of all durations than they do like children with completed spells of short duration. We would be much less interested in the characteristics of uninsured children by how long they have been uninsured were it not for the fact that state CHIP plans frequently limit coverage to children who have already accumulated some specified number of months of uninsurance. Table 22 compares the distributions of active spell lengths and completed spells by three characteristics: the child's age, family poverty level, and parents' employment.

TABLE 21: DISTRIBUTION OF SOCIOECONOMIC CHARACTERISTICS BY COMPLETED DURATION OF NEW SPELLS OF MEDICAID-ELIGIBLE UNINSURANCE, FY93
Socioeconomic Characteristic Completed Duration in Months
1 to 4 5 to 8 9 to 12 13+
Poverty Level 100 100 100 100
    Less than 50% FPL 25.4 31.3 18.1 20.7
    50% to < 100% FPL 34.5 34.9 37.9 30.9
    100% to < 200% FPL 34.8 28.6 38.1 39
    200% to < 300% FPL 2.8 4.5 1.4 5.3
    300% FPL or Greater 2.5 0.6 4.4 4
Parents' Employment 100 100 100 100
    1 or More Full Time 62.6 55.2 40.8 54.6
    Part Time Only 13.2 12 19.4 9.1
    No Working Parent 21.2 28.3 38.7 31
    No Parent Present 3 4.5 1.1 5.3
Parents' Education 100 100 100 100
    No Parent Present 3 4.5 1.1 5.3
    6 Years or Less 4.5 7.5 16.8 3.5
    7 to 11 Years 20.3 17.7 21.7 33.1
    12 Years 41.7 45.6 39.5 37.9
    Attended College 19.6 15.7 6.9 13.8
    4-Year College Degree 5.7 2.6 13.2 6.5
    Graduate Work 5.3 6.5 0.9 0
SOURCE: Survey of Income and Program Participation, 1992 Panel.

Active spells are limited to those of children who were uninsured in September 1993 while the completed spells represent all spells that were started in FY93.

The active spells include a higher proportion of very long spells than do the completed spells; 46 percent of the children uninsured in September 1993 had already been uninsured for 13 months or more whereas only 20 percent of all the spells that started in FY93 (which may include multiple spells by the same individuals) extended beyond 12 months.

Infants have much shorter active spells than completed spells because spell lengths are bounded by their very limited life spans. We see no infants with spells exceeding 12 months, of course, because no infant, by definition, has been alive for more than 12 months. By contrast, there is no limit to the completed spell length of children who were infants when their spells began, and infants have nearly the same proportion of completed new spells exceeding 12 months as older children: 17 percent compared to 18 to 22 pecent for older children). The proportion of older children whose active spells exceed 12 months rises 20 percentage points--from 36 percent to 56 percent--between ages 1 to 5 and ages 11 to 15 but then drops slightly, to 49 percent, in the next age group. The proportion of new spells exceeding 12 months in length shows no increase with age after infancy, however. It is not clear why we should see an age difference in the current duration of active spells when there is no age difference in the completed durations of new spells. Active spells are in some sense sampled from new spells but with a probability proportional to (completed) length. It must be that the age differences that we see among active spells are due to age differences that do not appear in new spells until durations well beyond 13 months.

TABLE 22: CURRENT DURATION OF ACTIVE SPELLS COMPARED TO COMPLETED DURATION OF NEW SPELLS BY SELECTED CHARACTERISTICS: SPELLS ACTIVE IN SEPTEMBER 1993 AND NEW SPELLS BEGINNING IN FY93
Characteristic Current Durationof Active Spells in Months Completed Duration of New Spells in Months
1 to 4 5 to 8 9 to 12 13+ Total 1 to 4 5 to 8 9 to 12 13+ Total
All Children 34.6 12.1 7.5 45.8 100 54 16.9 9 20.1 100
Age of Child                    
    Infant (0) 81.3 12.5 6.2 0 100 58.9 15.5 8.2 17.4 100
    1 to 5 37.8 13.9 12.1 36.3 100 56.7 14.2 9.3 19.8 100
    6 to 10 36.4 12 6.2 45.4 100 51.4 18.8 8 21.7 100
    11 to 15 28.4 10.3 5.3 56 100 55.1 16.3 10.4 18.2 100
    16 to 18 30.8 12.6 7.5 49.2 100 49.8 20.2 8.1 21.8 100
Poverty Level                    
    Less than 50% FPL 37.3 10.2 7.6 44.8 100 62.8 15.5 8.1 13.7 100
    50% to < 100% FPL 39 11 6.3 43.7 100 55.3 17 8.6 19.1 100
    100% to < 200% FPL 31.2 11.3 7.6 49.8 100 50.8 17.8 8.6 22.8 100
    200% to < 300% FPL 31.8 14.5 8.7 45 100 51.2 13.6 11.4 23.8 100
    300% FPL or Greater 40.7 16.4 7.8 35.1 100 53 19.7 9.5 17.8 100
Parents' Employment                    
    1 or More Full Time 33.6 11.9 7.6 46.9 100 53.5 17.2 8.5 20.8 100
    Part Time Only 35.7 13 6 45.3 100 49.8 20.2 13.1 16.8 100
    No Working Parent 39.6 11 5.8 43.6 100 56.3 14.5 10.5 18.7 100
    No Parent Present 34.2 14.1 15.1 36.6 100 60.3 15.6 2.6 21.5 100

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: For active spells, characteristics refer to September 1993. For new spells, characteristics refer to the start of the spell.

There are differences in the durations of both completed spells and active spells by poverty level, but the differences among active spells are weaker. The proportion of children with completed spells of 1 to 4 months declines from 63 percent to 51 percent over the first three income classes while the change among active spells is only 6 percentage points. Similarly, the proportion of new spells exceeding 12 months in length rises from 14 percent to 23 percent over the first three income classes whereas the increase among active spells is only 5 percentage points. Both distributions show an upturn in the relative frequency of short spells as income rises to more than 300 percent of poverty.

Parents' employment shows little relationship to the duration of completed spells. Children with no parent present have a somewhat higher frequency of short spells than other children, but there are no other consistent differences. Among children with active spells, those with no parents present have fewer spells exceeding 12 months than other children but comparable numbers of spells exceeding 8 months. Children with no working parents are somewhat more likely to have been uninsured for less than 5 months than children with working parents, but again the differences are modest.

Differences in the shares of active spells by demographic and socioeconomic groups are of interest because of what they tell us about the composition of children who may be eligible for state CHIP coverage. As we have pointed out, infants by definition cannot be uninsured for more than 12 months. States that would like to cover infants but choose to limit their coverage to children who have been uninsured for 12 months or more or even 6 months or more will have to consider defining eligibility among infants on some basis other than the length of their own spells of uninsurance, or infants will receive disproportionately little coverage. In Table 23 we see that while infants account for 7 percent of the spells of uninsurance that started 1 to 4 months earlier, they account for only 3 percent of the spells that started 5 to 8 months earlier and closer to 2 percent of the spells that started 9 to 12 months earlier. Shares of active spells that started more than 12 month earlier rise with age.

Children 1 to 5 account for 18 percent of such spells while children 11 to 15 account for 35 percent. There is little difference, however, in the age group shares of spells lasting 9 months or longer.

Both active spells and completed spells show the share of children in families between 100 and 200 percent of poverty rising with duration. This group accounts for 36 percent of the active spells that started 1 to 4 months earlier compared to 44 percent of the spells that started more than 12 months earlier. Completed spells show the share of children in this group rising from 34 percent to 41 percent. Other poverty classes show less variation in their shares of children with different durations of uninsurance.

Shares of the uninsured by parents' employment show minimal variation by parents' employment. For active spells, no employment group's share varies by more than 4 percentage points. There is somewhat greater variation among completed spells, but no group shows consistent growth or decline in its share of all spells as duration increases.

On the whole, then, spells active at a point in time show no greater evidence of demographic or socioeconomic differentials in duration than new spells that started over the course of a year.

TABLE 23: DISTRIBUTION OF SELECTED CHARACTERISTICS BY SPELL LENGTH: SPELLS ACTIVE IN SEPTEMBER 1993 AND NEW SPELLS BEGINNING IN FY93
Characteristic Currenthurationof Active Spells Completed Duration of New SpSpells
1 to 4 5 to 8 9 to 12 13+ 1 to 4 5 to 8 9 to 12 13+
Age of Child 100 100 100 100 100 100 100 100
    Infant (0) 6.8 3 2.4 0 6.9 5.8 5.8 5.5
    1 to 5 24.9 26.3 36.7 18.1 29.8 23.7 29.4 28.1
    6 to 10 27.3 25.9 21.3 25.7 23.1 27 21.7 26.2
    11 to 15 23.8 24.7 20.4 35.4 24.9 23.5 28.1 22.1
    16 to 18 17.2 20.1 19.2 20.8 15.4 19.9 15 18.1
Poverty Level 100 100 100 100 100 100 100 100
    Less than 50% FPL 13.6 10.7 12.7 12.3 17.8 14 13.8 10.4
    50% to < 100% FPL 25 20.2 18.6 21.1 24.5 24 22.8 22.7
    100% to < 200% FPL 36.5 38 41 44 34.3 38.3 34.9 41.4
    200% to < 300% FPL 15.1 19.7 18.9 16.1 12.2 10.3 16.3 15.3
    300% FPL or Greater 9.9 11.4 8.8 6.5 11.3 13.4 12.2 10.2
Parents' Employment 100 100 100 100 100 100 100 100
    1 or More Full Time 67.3 68.4 70.6 70.9 65.8 67.6 62.5 68.9
    Part Time Only 15.7 16.3 12.1 15 9.1 11.8 14.5 8.3
    No Working Parent 12.8 10.2 8.7 10.6 19.2 15.8 21.5 17.1
    No Parent Present 4.3 5 8.7 3.4 5.9 4.9 1.5 5.6

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: Active spell characteristics refer to September 1993. New spell characteristics refer to the start of the spell

D. Annual Incidence and Duration of Uninsurance, Medicaid Eligibility, and Medicaid Participation by Age

When we measure age at a point in time but measure behavior over a period of time, some of the behavior that we observe may have happened at an earlier or later age. If the behavior in question is affected by crossing particular age boundaries--as is Medicaid participation--then the relationship that we observe between age and the measured behavior may be attenuated by defining age at a point in time. Tables 24 through 27 are based on a different approach to assigning age to behavior measured over time. These tables examine the experience of uninsurance, Medicaid participation, and both Medicaid-eligible and Medicaid-ineligible periods of uninsurance through the 12 months that a child spends at a given single year of age. To create these tables, we identified children by their age in September 1993 and then looked forward and backward to determine the 12-month period that each child spent at the defined age.11 We then counted the number of months over this 12-month period that each child was observed in each of four statuses: uninsured (Table 24), uninsured but eligible for Medicaid (Table 25), uninsured but not eligible for Medicaid (Table 26), and enrolled in Medicaid (Table 27). For each of these four statuses we tabulated by single year of age:

  • The proportion of children who experienced any months in that status during the year
  • The mean number of months in that status among those who spent 1 to 12 months
  • The proportion who were in that status for all 12 months, expressed as
    (1) a proportion of children who spent any time in that status during the year,
    (2) a proportion of all children
TABLE 24: ANNUAL FREQUENCY OF UNINSURANCE BY SINGLE YEAR OF AGE
Age of Child in September 1993 Number of Children Children Ever Uninsured  
Percentage Ever Uninsured Average Number of Months Uninsured Percentage Uninsured All 12 Months Percentage of All Children Uninsured 12 Months
All Children 70,868,000 21.7 7.01 29.1 6.3
0 3,469,000 16.4 4.87 9.6 1.6
1 3,008,000 18.4 6.39 21.9 4
2 4,558,000 19.7 6.7 25.4 5
3 4,043,000 18.5 6.21 19.9 3.7
4 4,004,000 20.3 6.49 22.5 4.6
5 3,867,000 19.8 6.81 30.2 6
6 3,993,000 21.9 6.47 21.7 4.8
7 3,607,000 19.3 6.6 24.7 4.8
8 3,841,000 20.9 6.74 29.7 6.2
9 3,532,000 21.7 7.17 28.7 6.2
10 3,785,000 22.2 7.05 29.6 6.6
11 3,742,000 23.9 7.45 32.7 7.8
12 3,931,000 23.4 6.67 26.9 6.3
13 3,906,000 23 8.28 39.1 9
14 3,461,000 21 7.9 35 7.4
15 3,376,000 23.8 7.63 37.8 9
16 3,578,000 25.3 7.55 38 9.6
17 3,678,000 25.9 7.33 29.8 7.7
18 3,490,000 26.9 7.73 38.1 10.2

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: Incidence and frequency of uninsurance were observed over the entire 12 months that
a child spent at the indicated age.

TABLE 25: ANNUAL FREQUENCY OF UNINSURANCE WITH MEDICAID ELIGIBILITY BY SINGLE YEAR OF AGE
Age of Child in September 1993 Number of Children Percentage Ever Uninsured and Medicaid- Eligible Children Ever Uninsured and Medicaid-Eligible Percentage of All Children Uninsured and Eligible 12 Months
Average Number of Months Percentage Uninsured and Eligible 12 Months
All Children 70,868,000 10.7 4.78 9 1
0 3,469,000 13 4.22 5.5 0.7
1 3,008,000 12.7 5.36 11.1 1.4
2 4,558,000 14.1 5.4 9.4 1.3
3 4,043,000 12.7 5.41 9.8 1.2
4 4,004,000 14.7 4.56 6.9 1
5 3,867,000 15.3 5.4 12.3 1.9
6 3,993,000 13.9 3.94 2.3 0.3
7 3,607,000 10.5 4.74 7.1 0.7
8 3,841,000 12.1 5.29 11.3 1.4
9 3,532,000 13 4.89 7.6 1
10 3,785,000 7.3 4.08 6.2 0.5
11 3,742,000 8.5 3.86 5.6 0.5
12 3,931,000 7.3 3.86 8.1 0.6
13 3,906,000 8.1 4.24 11.8 1
14 3,461,000 7.1 6.77 22.2 1.6
15 3,376,000 9.5 4.09 4.1 0.4
16 3,578,000 7.7 4.69 10.6 0.8
17 3,678,000 7.6 4.01 6.9 0.5
18 3,490,000 6.4 5.68 24.7 1.6

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: Incidence and frequency of uninsurance were observed over the entire 12 months that
a child spent at the indicated age.

TABLE 26: ANNUAL FREQUENCY OF UNINSURANCE WITHOUT MEDICAID ELIGIBILITYBY SINGLE YEAR OF AGE
Age of Child in September 1993 Number of Children Percentage  Ever Uninsured and Not Medicaid- Eligible Children Ever Uninsured and Not Medicaid-Eligible
 
Percentage of All Children Uninsured and Ineligible 12 Months
Average Number of Months Percentage Uninsured and Ineligible 12 Months
All Children 70868000 16.3 6.19 20.1 3.3
0 3469000 6 4.1 5 0.3
1 3008000 11.6 4.27 1.5 0.2
2 4558000 10.6 5.27 12.2 1.3
3 4043000 11.8 3.91 4.4 0.5
4 4004000 12.2 5.33 14.7 1.8
5 3867000 11.6 4.48 10.4 1.2
6 3993000 15.3 5.69 12.6 1.9
7 3607000 14 5.52 12.6 1.8
8 3841000 13.9 5.52 14.8 2.1
9 3532000 17.1 5.39 11 1.9
10 3785000 19.6 6.46 24.2 4.7
11 3742000 20.4 7.12 30.1 6.2
12 3931000 20 6.4 23.6 4.7
13 3906000 20.2 7.72 29.6 6
14 3461000 17.1 6.91 23.7 4.1
15 3376000 19.8 7.19 31.4 6.2
16 3578000 22.9 6.77 26.3 6
17 3678000 23.4 6.8 19.1 4.5
18 3490000 23.7 7.26 32.5 7.7

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: Incidence and frequency of uninsurance were observed over the entire 12 months that
a child spent at the indicated age.

TABLE 27: ANNUAL FREQUENCY OF MEDICAID COVERAGE BY SINGLE YEAR OF AGE
Age of Child in September 1993 Number of Children Percentage Ever Covered by Medicaid Children Ever Covered by Medicaid Percentage of All Children Covered for All 12 Months
Average Number of Months Percentage Covered for All 12 Months
All Children 70868000 24.2 9.21 55.1 13.3
0 3469000 39.2 10.28 63.6 24.9
1 3008000 34.7 9.61 61.5 21.3
2 4558000 35.4 9.67 62.7 22.2
3 4043000 29.4 9.41 60.8 17.8
4 4004000 27.8 9.52 53.5 14.8
5 3867000 28.4 9.95 65.2 18.5
6 3993000 30.3 8.97 50.5 15.3
7 3607000 21.5 9.55 58.9 12.7
8 3841000 22.2 9.54 60.7 13.5
9 3532000 24.8 8.83 52.6 13
10 3785000 20.9 8.79 49.8 10.4
11 3742000 20.9 8.81 51.4 10.7
12 3931000 19.5 8.47 45.4 8.9
13 3906000 20.3 8.84 55.6 11.3
14 3461000 19 8.23 44.6 8.5
15 3376000 17.1 8.81 47.5 8.1
16 3578000 15.2 8.91 48.3 7.3
17 3678000 19.3 8.04 42.4 8.2
18 3490000 12.2 7.98 33.4 4.1

SOURCE: Survey of Income and Program Participation, 1992 Panel.

NOTE: Incidence and frequency of Medicaid coverage were observed over the entire 12 months that a child spent at the indicated age.

These tabulations give us a clearer picture of the age-specific patterns of uninsurance and Medicaid participation than we obtain when we classify children's behavior over the course of a year by their age at the beginning or end of that year.

1. Uninsurance

In Table 24 we see that among infants in September 1993, 16.4 percent were ever uninsured during their infancy. Among those who were ever uninsured the average number of months without insurance was 4.87, and 9.6 percent of these children were uninsured for all 12 months. The proportion of all infants who were uninsured for the entire year of their infancy was 1.6 percent. From the average duration of uninsurance we could calculate the probability that an infant who was ever uninsured during the year was uninsured at a point in time. This is simply the average number of months uninsured divided by 12. Multiplying this probability (.406) by the percentage ever uninsured during the year gives us the average percentage of all infants--6.7 percent--who would have been uninsured at a point in time. For 18-year-olds this same calculation would yield 17.3 percent as the proportion likely to have been uninsured at a point in time.

The percentage of children ever uninsured during the year shows a progressive rise from age 0 to 18, reaching 26.9 percent among 18-year-olds. The average duration of uninsurance rises as well but not as steeply as the percentage of ever-uninsured children who were uninsured for the entire year. This latter approaches 40 percent among the oldest children whereas it is less than 10 percent among infants and just over 20 percent in the next higher ages. (These estimates show marked variability among neighboring ages, which we attribute primarily to sampling error.) The percentage of all children uninsured for 12 months rises even more dramatically because it is the product of two percentages that rise with age: the percentage ever uninsured and the percentage among these who were uninsured the entire year. From a low of 1.6 percent among infants and 4.0 percent among children one year of age the proportion uninsured for all 12 months rises to just over 10 percent among children age 18.

Turning to Medicaid eligibility within spells of uninsurance, we find very different patterns by age than we do for uninsurance generally. In Table 25 we see that just under 11 percent of all children were ever Medicaid-eligible and uninsured during the year. This fraction is 13 percent among infants, and it remains at about that level until age 10, when it drops to between 7 and 8 percent. Eligibility under the poverty-related criteria did not extend to children who were age 10 in September 1993; children 10 and older at that point could qualify for Medicaid through a number of other channels, but the unavailability of coverage through the poverty-related criteria limited eligibility to barely more than half the number who were eligible at younger ages.

That the probability of being both uninsured and Medicaid-eligible varies so little by age can be explained by the fact that Medicaid eligibility rates decline with age while the probability of being uninsured rises. With these two rates moving in opposite directions their product tends to remain nearly constant. By the same reasoning, however, we would expect the probability of being uninsured and not eligible for Medicaid to rise sharply with age. Both the probability of being uninsured and the probability of being ineligible for Medicaid increase with age. Hence the age pattern will be amplified when these two statuses are combined.

The average number of months that children were uninsured and Medicaid-eligible did not vary by age, it appears.12 Children remained in this state for an average of nearly five months at every year of age. We suggest that the uniformity of this experience by age may be due to the fact that, unlike spells of uninsurance without Medicaid eligibility, parents could choose to end a spell of Medicaid-eligible uninsurance by enrolling the child in Medicaid. There is no obvious reason why the age of the child should affect how long parents choose to wait, except perhaps for infants, although spells for infants were no shorter than spells for older children.

The proportion of ever-uninsured (and Medicaid-eligible) children who were uninsured for all 12 months shows very considerable sampling error. This proportion seems to rise with age whereas the percentage of all children who were uninsured and Medicaid-eligible for all 12 months remains steady or even falls.13

Periods of uninsurance without Medicaid eligibility occurred with a frequency between that of Medicaid-eligible uninsurance and all uninsurance (Table 26). Because uninsured children below age 10 were nearly twice as likely to be eligible for Medicaid as children 10 and older, the proportion of children who were ever uninsured and not eligible for Medicaid during the year rises more sharply with age than the percentage who were simply ever uninsured. From 6 percent among infants the percentage ever uninsured and not Medicaid-eligible rises to 24 percent among 18-year-olds. Mean durations in this state rose by two to three months over the range of ages, with an average duration of 6.2 months, while the percentage of those remaining in the state for 12 months grew from the low single digits to over 30 percent, with an average of 20 percent. Among all children, the percentage who were uninsured without Medicaid eligibility for the entire year rose from .3 percent among infants to nearly 8 percent among 18-year-olds.

2. Medicaid Participation

Table 27 presents statistics on reported Medicaid coverage or participation by single year of age. Among all children, 24.2 percent were ever covered by Medicaid during the year. Those who were ever covered were covered for an average of over 9 months, and 55.1 percent were covered for the entire year. Among all children the percentage covered for all 12 months was 13.3 percent.

There is a steep age gradient in Medicaid participation, but it does not show the clear break between ages 9 and 10 that we see in the tabulations of simulated eligibility. Participation is highest among infants, with 39 percent ever covered during the year. Infants with coverage are enrolled for an average of more than 10 months of their infancy, with about 64 percent of this group or 25 percent of all infants covered for the entire year.14 The percentage of children ever covered during the year drops to 35 percent for ages one and two and then falls to between 28 and 30 percent through age six. From about 25 percent at age nine the proportion ever covered during the year descends to 21 percent at age 10, then declines gradually to 19 percent by age 14, drops to 15 percent by age 16, rises to 19 percent, and then plunges to 12 percent at age 18.

Among children ever covered by Medicaid, both the average number of months of coverage during the year and the percentage covered for all 12 months decline with age. The average number of months of coverage drops from 10.3 months for infants to 8 months by age 18. The percentage of children covered for all 12 months falls from 64 percent among infants to 33 percent at age 18. Among all children, the percentage covered by Medicaid for all 12 months of the year shows an even more dramatic decline, from 25 percent among infants to just 4 percent among 18-year-olds.

E. Conclusion

This report has used data from the 1992 panel of the SIPP to examine demographic and socioeconomic differentials in the patterns of health insurance coverage among children under 19. Health insurance coverage among children varies by nearly every demographic and socioeconomic characteristic that we examined. Most of the differentials that we observe in the type of insurance coverage and whether there is any coverage at all are moderately strong to very strong. For example, Hispanic children are more than two-and-a-half times as likely to be uninsured as white non-Hispanic children, and black children are four times as likely as white children to be covered by Medicaid.

Because of Medicaid, coverage patterns are not unidimensional. Groups with low rates of employer-sponsored coverage do not necessarily have high rates of uninsurance. High rates of Medicaid coverage can appear among groups with high uninsurance or moderately low rates of uninsurance. Nevertheless, the strongest differentials by far are those associated with parents' education, and these differentials are strikingly unidimensional.

We estimate that one-third of the children uninsured in September 1994 were eligible for Medicaid, based on a Medicaid eligibility simulation that takes account of most but not all of the ways that children may become eligible. There are sizable differentials in Medicaid eligibility across most of the demographic and socioeconomic characteristics that we examined. Differentials in Medicaid eligibility among the uninsured are strongest by family poverty level, followed by the child's age and parents' employment.

Over all eligible children, the estimated Medicaid participation rate is 65 percent. This participation rate does not reflect any adjustment for the underreporting of Medicaid participation in the SIPP, which in 1994 was on the order of 24 percent. If we exclude from the eligible nonparticipants those who were covered by insurance other than Medicaid, however, the participation rate rises to 79 percent. Differentials by demographic characteristics are largely eliminated when the participation rate is defined in this alternative way, and socioeconomic differentials are greatly reduced.

Generally, differentials in the proportion of children ever uninsured during the year are similar in form to differentials in the percentage uninsured at a point in time. Only for race and ethnicity and region did we find that groups were arrayed differently with respect to their likelihood of being uninsured. Distributions of the uninsured by demographic or socioeconomic characteristics differ very little between the point-in-time and annual-ever measures.

The duration of new spells of uninsurance shows little variation by demographic characteristics or parents' employment. There is modest variation in spell length by poverty level and moderately strong variation by parents' education, however. Spell lengths decline with parents' education.

The current duration of active, or ongoing, spells--that is, how long uninsured children have been without coverage--is of interest recently because a number of states are planning to limit eligibility under their CHIP initiatives to children who have been uninsured for some minimum number of months--as many as 12 months. It appears that such restrictions favor the group that is the principal target of CHIP: children between 100 and 200 percent of poverty. This is the one subgroup whose share of the uninsured clearly increases with duration.

We also examined the insurance coverage of children's parents. Between 18 and 21 percent of uninsured children have an insured parent. This suggests that a nontrivial number of parents may be choosing individual coverage but not family coverage--perhaps because family coverage is not offered or is perceived as too expensive. Measurement error may account for part of this finding as well. Parents may report their own coverage but overlook the coverage that extends to their children. This seems particularly likely for the 7 percent of uninsured children who report that a parent is covered by Medicaid. Among children who are themselves reported to be covered by Medicaid, 10 percent have an uninsured parent. This latter is consistent with the child-only eligibility created by the poverty-related expansions of Medicaid in the late 1980s and 1990s.

Finally, to provide detailed information on life cycle patterns of coverage among children, we estimated the frequency with which children experienced periods of uninsurance, with and without Medicaid-eligibility, during each single year of age. We also estimated the frequency with which children experienced periods of reported Medicaid enrollment. Both the probability of being uninsured and the average number of months uninsured among those with any months of uninsurance increase with age. The likelihood of being uninsured for all 12 months shows a particularly strong relationship with age. The probability of being uninsured and eligible for Medicaid has little relationship with age, however, being relatively constant from infancy through age 9 and then dropping a few percentage points to a level that remains essentially fixed through age 18. The average number of months of Medicaid-eligible uninsurance is relatively constant across the entire age range as well. By contrast, uninsurance without Medicaid eligibility is even more strongly associated with age than uninsurance in general. The estimates of Medicaid participation by single year of age make clear the prominent role of Medicaid as a source of coverage among infants and the increasingly smaller role that Medicaid plays as children grow older. As more children become eligible for the federally mandated coverage for children below the poverty line who were born after September 30, 1983, however, these age differentials will gradually diminish.

Endnotes

1. Each year the upper age limit for federally mandated coverage rises by one year, in effect.

2. Foster children may be covered by Medicaid and, depending on whose household they appear in the SIPP, could be reported as having no parents in the household.

3. The SIPP instrument includes questions on insurance coverage provided to household members by persons outside the survey household, but it is not difficult to imagine that such coverage is reported less completely than coverage provided by parents or other adults in the household.

4. The poverty thresholds, which are provided on the SIPP file, are the same thresholds that the Census Bureau used to calculate the official estimates of children in poverty in 1994.

5. Monthly poverty rates run higher than annual poverty rates, generally, but annual poverty rates obtained from the SIPP by aggregating monthly income and poverty thresholds for individuals tend to run lower than the official poverty rates estimated from the Current Population Survey (CPS). The difference in the annual estimates can be attributed in large part to the SIPP's more accurate attribution of income to family members actually present each month.

6. These estimates of Medicaid eligibility are based on a detailed simulation described in Technical Appendix A. The simulation uses monthly income (in this case for September 1994) and other characteristics--such as the child's age and family composition--measured in the same month, along with state-specific eligibility criteria. The use of monthly data yields a more accurate simulation of the actual Medicaid eligibility determination than does the use of annual data or characteristics measured only once during a year. Other things being equal, monthly income would yield more eligibles than annual income, but there are a number of factors that confound the comparison of our monthly simulation with what other researchers have done with annual data. Finally, our simulation does not include the spend down features of the medically needy program, because the SIPP does not collect medical expenditure data. Therefore we know that we understate eligibility for Medicaid. We are not aware of any other simulations that are comparable to ours in other respects but include this feature of eligibility.

7. We exclude children who were reported to be participating in Medicaid but were simulated to be ineligible. We do so because many of the simulated ineligible participants may have been eligible for Medicaid under provisions that we did not simulate. To include just the participants would be equivalent to assuming that there were no nonparticipants among this additional group of eligibles when the participation rates that we observe for children eligible under related provisions suggest that there may have been at least twice as many nonparticipants as participants.

8. If some of these children were actually enrolled in Medicaid and their coverage had simply been misreported, we would want to include them in both the numerator and denominator of the participation rate. If all of these children were actually enrolled in Medicaid, including them in this way would yield a Medicaid participation rate of nearly 83 percent.

9. It is plausible, as well, that some number of spells that lasted fewer than four months were not reported at all. On balance, we suspect that spells of 1 to 4 months in length are still overstated, but the potential exclusion of some spells altogether has implications for the estimated number of children who were uninsured at a point in time or ever in a year. In fact, all misreporting of durations may affect estimates of incidence as well..

10. Fewer than 5 percent of the uninsured children in families above 200 percent of poverty were simulated as Medicaid eligible.

11. For example, a child who was observed at age 1 in September 1993 could have turned 1 in any of the months from October 1992 through September 1993 and ended that year of age in any of the months from September 1993 through August 1994.

12. We attribute the variability that we see to sampling error.

13. Recall that the proportion eligible for all 12 months among all children is the product of the proportion eligible all 12 months among those ever eligible, which rises with age, and the proportion who are ever eligible, which declines with age.

14. The 39 percent of infants whom we find to have ever participated in Medicaid during their infancy and their average of 10.3 months of coverage imply that 33.5 percent of infants would have been covered by Medicaid at any point in time. This compares quite closely to the 34.2 percent who were reported as covered in September 1993 (Table B.3).

Technical Appendix D: How Many Uninsured Children Are There?

Introduction

The estimates presented in this series of technical appendices are accompanied by two caveats. First, they are based on a panel sample--that is, a sample of children who were followed over a (three-year) period rather than a sample of children observed at a single point in time. This has implications for what population the sample represents at any given point over the period that we observe it. Second, they indicate too few children covered by Medicaid, based on comparisons with Medicaid program statistics. Both of these caveats have implications for the overall number of children and the proportion of all children that we estimate to be uninsured at any one time or over any period of time. In this brief report we consider how our estimates of the uninsured might be changed if we could expand the representation of the SIPP panel and obtain complete reporting of Medicaid coverage. We conclude that complete reporting of Medicaid coverage would reduce the 1992 SIPP panel estimate of the percentage of children who were uninsured in September 1993 from 13.0 percent to between 10.2 and 10.8 percent and would reduce the estimated percentage of children who were uninsured in September 1994 from 12.6 percent to between 9.5 and 10.1 percent.

Representativeness of the SIPP Sample

At the time of the initial interview, the children in the SIPP sample represented the entire United States population of children except for those residing in institutions or military barracks.1 Over time, however, as the U.S. population of children under 19 grew to 74.0 million, the population of children represented by the SIPP longitudinal sample declined to less than 70 million. As we detailed in Technical Appendix A, except for births the SIPP sample does not represent children who entered the population after the initial interview--that is, children who migrated to the U.S., returned from abroad, or left institutions. Moreover, while the SIPP sample theoretically represents all children born to members of the initial population who remained in the U.S., we demonstrated that these births are underrepresented by as much as a quarter. It is this shortfall of sample births that accounts for the sizeable decline in the population of children that the SIPP sample represents.

The fact that most of the divergence between the SIPP population and the total U.S. population of children is concentrated in the youngest ages carries important implications. As we have seen, the health insurance coverage patterns of very young children are strikingly different from those of other children. Underrepresenting births will very likely affect the overall distribution of coverage, increasing the proportion uninsured and reducing the proportion covered by Medicaid

The fact that the SIPP sample does not represent immigration has implications for estimates of insurance coverage as well. We would expect the rates of insurance coverage to be lower among immigrant children than among the general population of children. Therefore, excluding the immigrants will lower both the absolute number and the proportion of children who are estimated to be without insurance relative to what we would observe with a sample of the total population at the same point in time.

In order to evaluate our SIPP estimates of Medicaid enrollees against those reported by the Health Care Financing Administration(now known as Centers for Medicare and Medicaid Services(CMS)) (HCFA(now known as CMS)), we must expand our estimates of Medicaid coverage to the entire population. This entails first estimating the number of children that we are "missing" due to the SIPP panel's underrepresentation of births and its non-representation of other additions to the population and then estimating the frequency of Medicaid coverage and lack of insurance among these additional children. Our derivation of adjusted population totals, estimates of uninsured children, and estimates of children enrolled in Medicaid is summarized in Table 1.

Year and Estimate of Year at End Under 19 Children in Year Uninsured Ever Number of Year at End Uninsured Number in Year Medicaid Enrolled in Number Ever of Year at End Medicaid Enrolled in
TABLE 1: ADDITIONS TO THE PANEL ESTIMATES OF CHILDREN EVER UNINSURED AND CHILDREN EVER ENROLLED IN MEDICAID, FY93 AND FY94
FY93          
Panel Estimate 70,868,000 16,089,000 9,271,000 17,800,000 13,369,000
Additions to Correct for Underrepresentation of Births 1,150,000 169,000 75,000 441,000 382,000
Additions to Include Other New Entrants to the Population 1,119,000 241,000 147,000 266,000 211,000
Adjusted Total 73,137,000 16,499,000 9,493,000 18,507,000 13,962,000
FY94          
Panel Estimate 69,935,000 15,936,000 8,911,000 17,795,000 13,259,000
Additions to Correct for Underrepresentation of Births 2,250,000 360,000 181,000 843,000 716,000
Additions to Include Other New Entrants to the Population 1,772,000 383,000 225,000 429,000 337,000
Adjusted Total 73,957,000 16,679,000 9,317,000 19,067,000 14,312,000
SOURCE: Panel estimates are taken from Technical Appendix A, Tables 1 and 4. Adjusted total populations are also from Table 1. Estimates of the proportion of infants and young children uninsured or enrolled in Medicaid are based on Technical Appendix C, Tables 24 and 27. The estimates of the proportion of children uninsured or enrolled in Medicaid that were applied to new entrants were taken from Technical Appendix A, Table 4.

For September 1993 we added 1.150 million children to correct for the underrepresentation of births and another 1.119 children to account for net growth of the population of children. For September 1994 we added 2.250 million children to correct for the underrepresentation of births and another 1.772 million children to incorporate other new entrants to the population.

To derive estimates of uninsured children and Medicaid enrollees among the children added to correct for underrepresented births, we applied age-specific rates from two tables that were presented in Technical Appendix C. To derive estimates of uninsured children and Medicaid enrollees among the other new entrants to the population, we assumed that these children shared the same probabilities of being uninsured or covered by Medicaid as all children in 1993 and 1994, and we applied rates from a table presented in Technical Appendix A. This is not an entirely satisfactory assumption, given that these new entrants to the population could differ dramatically from the rest of the population. Nevertheless, in the absence of relevant data on this population, the assumption is a reasonable one to apply in this case, and the final results are not very sensitive to it.

The net effect of these operations is to add about 700,000 children to the number of children estimated to have been ever enrolled in Medicaid during FY93 and 1.27 million to the number estimated to have been ever enrolled in Medicaid in FY94. Children added to correct for the underrepresentation of births accounted for about two-thirds of the additional Medicaid enrollees in each year.

Undercount of Medicaid Children

Crude comparisons of CPS estimates and Medicaid program statistics suggest that the CPS underestimates by a substantial margin the number of Medicaid enrollees for the population as a whole. Researchers at the Urban Institute have made a practice of adjusting their estimates of uninsured children and adults to compensate for what they interpret as incomplete reporting of Medicaid coverage by survey households. The effect of these adjustments is sizable. When applied to data for 1994 and 1995 they lowered the estimated number of uninsured children from 10.6 million to 7.6 million and reduced the estimated number of Medicaid-eligible uninsured children from 4.5 million to 1.6 million (Ullman et al. 1998).

Evaluating the Medicaid coverage of the SIPP or the CPS is, in fact, both difficult and problematic. It is difficult because the program administrative statistics are very limited, which complicates the task of creating administrative and survey estimates that apply to the same universe. One major problem, discussed in subsequent sections, is that survey data generally represent points in time while federal administrative data include those covered by Medicaid at any time throughout the year. Beyond that, the age categories for which Medicaid statistics are reported do not lend themselves to identifying children in the way that they are commonly defined in policy analyses--that is, under 19 or under 18. Part of the problem is that the national statistics are assembled from reports submitted by the states, and the state data differ in quality and in some of the conventions that they apply. In 1993 more than half of the state reports were submitted in hard-copy, giving HCFA(now known as CMS) no flexibility to generate additional tabulations--with alternative age categories, for exmple.

Evaluating Medicaid coverage is also problematic. Even if the program data and survey data could be matched with respect to universe, simply comparing them would not tell the whole story. Because the Medicaid statistics come from separate state systems they are not unduplicated across states. If a child moves from one state to another and is enrolled in Medicaid in both states during the same fiscal year, the child will be counted in both states' statistics. In a survey, of course, this same child would be represented as a Medicaid recipient in only one state. Another factor confounding the comparison of survey and administrative estimates is the potential misreporting of Medicaid coverage as something else. Survey respondents may not be fully aware that the source of their medical coverage is Medicaid. States have developed alternative names for their programs and issued health insurance cards that resemble those issued by private insurers, perhaps persuading some Medicaid enrollees that their coverage is from a private source. Finally, a number of states run affiliated programs that do not receive federal reimbursement and whose enrollees, therefore, are not counted in the Medicaid program statistics. If such persons report their coverage in surveys as Medicaid, as seems likely, or if the Census Bureau interprets it and edits it as such, then the survey estimates of Medicaid coverage will include persons who are not counted in the administrative totals.

Impact of the Medicaid Undercount on Estimates of the Uninsured

Determining how the Medicaid undercount affects estimates of the uninsured is even more problematic. Apart from the issue of whether Medicaid coverage may be reported as something else is the matter of other coverage in general. If a child did not report Medicaid and did not misreport it as something else, the child might still have had--and reported--another source of coverage.2 To estimate the impact of even a known Medicaid undercount on the number of children who are estimated to be uninsured requires assumptions about how the children who failed to report their Medicaid coverage may have reported their insurance status with respect to other potential sources of coverage.

Estimates of the Medicaid Undercount

Based on a comparison of the adjusted Medicaid enrollment estimates in Table 1 with estimates derived from program statistics, we determined that the undercount of Medicaid enrollees in the 1992 SIPP panel was 13.2 percent or 2.8 million children in FY93 and 14.6 percent or 3.3 million children in FY94 (Table 2).3

Attributing all of the difference between the survey and administrative estimates to a Medicaid undercount presumes that the administrative count is in fact accurate. While we have reason to believe that there is some double-counting of Medicaid enrollees, HCFA(now known as CMS) has advised users of its program data that participating infants may be undercounted by about 10 percent in as many as half the states (HCFA(now known as CMS) 1994). Furthermore, we have not attempted to estimate the number of children enrolled in fully state-funded programs. These children are not included in the statistics reported to HCFA(now known as CMS) by the states, but they are included, to at least some degree, in the survey estimates of the Medicaid population. On balance, then, we have assumed that these errors and omissions in the administrative statistics net to zero.

Estimates of Uninsured Children

To estimate the impact of the Medicaid undercount on estimates of the number of uninsured children, we must work with estimates of children uninsured at a point in time rather than children ever uninsured during the year. Having Medicaid (or any other form of coverage) during a year does not imply that a child was insured throughout the entire year. Therefore, the impact of a Medicaid undercount on estimates of the number of children who were ever uninsured is too ambiguous to estimate.

Estimate FY93 FY94
TABLE 2: COMPARISON OF SIPP AND HCFA(now known as CMS) ESTIMATES OF CHILDREN EVER ENROLLED IN MEDICAID, FY93 AND FY94
Adjusted SIPP Panel Estimate 18,507,000 19,067,000
Age-adjusted HCFA(now known as CMS) Estimate 21,317,000 22,324,000
SIPP Undercount 2,810,000 3,257,000
Estimated Undercount as Percent of HCFA(now known as CMS) Estimate 13.2% 14.6%
SOURCE: Survey of Income and Program Participation, 1992 Panel, and HCFA(now known as CMS) 2082 Reports.

Our estimate of the Medicaid undercount represents children who were ever enrolled in Medicaid during a year. While the proportionate undercount need not be the same for the annual-ever and point-in-time statistics, we assumed that it was. In Table 3 we derive estimates of the size of the SIPP Medicaid undercount for September 1993 and 1994.4 We obtained 2.120 million as the estimated number of children who were covered but did not report Medicaid enrollment in September 1993 and 2.445 million as the number who were covered but failed to report their Medicaid enrollment in September 1994.

What proportion of the children who failed to report their Medicaid enrollment also failed to report any other kind of coverage? As we discussed earlier, some children (or their parents, in most cases) may have misreported their Medicaid coverage as another form of public or even private insurance. Others may have even had overlapping coverage and reported the other coverage rather than Medicaid.5 With no evidence to support specific assumptions about these proportions, we applied three alternative assumptions. We assumed that the proportion of unreported Medicaid enrollees who were classified as uninsured was either 95 percent, 85 percent, or 75 percent. These different assumptions may be considered to include our uncertainty about the incidence of duplication of individuals in the national Medicaid statistics.

Estimate   September 1993 September 1994
TABLE 3: IMPACT OF MEDICAID UNDERCOUNT ON THE ESTIMATED NUMBER OF CHILDREN UNDER 19 WHO WERE UNINSURED IN SEPTEMBER 1993 AND 1994
Adjusted SIPP Panel Estimates    
Number of Children Uninsured 9,493,000 9,317,000
Percentage of Children Uninsured 13.0% 12.6%
Number of Children Enrolled in Medicaid 13,962,000 14,312,000
     
Estimated Percentage of Medicaid Children Not Counted in SIPP 13.2% 14.6%
     
Undercount of Children Who Were Enrolled in Medicaid in September 2,120,000 2,445,000
     
Assumed Proportion of Unreported Medicaid Enrollees Counted as Uninsured    
Alternative 1 95% 95%
Alternative 2 85% 85%
Alternative 3 75% 75%
     
Estimated Number of Unreported Medicaid Enrollees Counted as Uninsured    
Alternative 1 2,014,000 2,323,000
Alternative 2 1,802,000 2,078,000
Alternative 3 1,590,000 1,834,000
     
Revised Estimate of the Number of Children Uninsured in September    
Alternative 1 7,479,000 6,994,000
Alternative 2 7,691,000 7,239,000
Alternative 3 7,903,000 7,483,000
     
Revised Estimate of the Proportion of Children Uninsured in September    
Alternative 1 10.2% 9.5%
Alternative 2 10.5% 9.8%
Alternative 3 10.8% 10.1%
     
SOURCE: Survey of Income and Program Participation, 1992 Panel.

The alternative assumptions imply that between 1.6 million and 2.0 million children were incorrectly counted as uninsured rather than insured in September 1993, and they lower the estimated number of uninsured children to between 7.5 million and 7.9 million--compared to the SIPP panel estimate of 9.5 million. Taking into account the Medicaid undercount implies that between 10.2 and 10.8 percent of children under 19 were uninsured in September 1993, compared to the SIPP panel estimate of 13.0 percent. For September 1994 the higher estimated Medicaid undercount yields estimates of the number of uninsured children that range between 7.0 million and 7.5 million compared to the SIPP panel estimate of 9.3 million. The alternative estimates of the proportion of children without health insurance range from 9.5 percent to 10.2 percent versus the SIPP panel estimate of 12.6 percent.

Endnotes

1. As a household-based survey, the SIPP also excludes the homeless, although homeless children are included in the population totals to which the SIPP sample weights are controlled. In this sense, then, the homeless are counted in the population estimates that we present, but their characteristics are not represented. If homeless children are more likely to be uninsured or more likely to be enrolled in Medicaid than children in the household population, for example, the frequency of these characteristics in the total population will be understated.

2. While we speak as if the child is the one who gave the responses, the CPS and SIPP collect data on children under 15 from other members of the household.

3. The HCFA(now known as CMS) data are reported for groups of ages, with ages 15 to 20 included in a single category. Since our SIPP estimates apply to children under 19, we had to estimate the proportion of children 15 to 20 who were under 19. We tabulated the March 1994 CPS to obtain the distribution of Medicaid enrollees under age 21 by single year of age, and based on these data we apportioned the HCFA(now known as CMS) count of enrollees 15 to 20 into ages 15 to 18 and ages 19 to 20. States use at least two different conventions for reporting age to HCFA(now known as CMS). States that participate in the Medicaid Statistical Information System (MSIS) and submit electronic case records to HCFA(now known as CMS) assign age as of the end of the fiscal year. States that submit 2082 reports instead of MSIS electronic case records assign age as of the middle of the fiscal year. The latter system classifies about 50 percent more children as infants and reports more persons under any given age than does the first system. (States using the first system may also exclude children born in the final month or two of the fiscal year, according to HCFA(now known as CMS).) Depending on whether a state appeared to have a substantial excess of infants or not, we applied a different algorithm to estimate the number of Medicaid enrollees who were under 19 in each state and then aggregated these to obtain the national numbers shown in Table 2.

4. To obtain the estimated undercount for 1993, we divided the adjusted SIPP estimate of children enrolled in Medicaid in September 1993 by one minus the percentage undercount to obtain the implied total Medicaid enrollment in that month, from which we then subtracted the adjusted SIPP estimate.

5. In assigning insurance coverage each month, we classified a child as enrolled in Medicaid if Medicaid was reported in combination with any other type of coverage.

Populations
Children