Formal and Informal Kinship Care Arrangements
Up to this point, the kinship care relationship has been described as a single category. However, one of the main reasons kinship care has drawn interest from policymakers and program administrators is the recent growth of one particular subset of kinship-caregiving relations -- kinship foster care, or formal kinship care. The emergence of formal kinship care as an important policy topic gives rise to many questions. Are kinship arrangements that are formally sanctioned and supported by state child welfare systems fundamentally different from informal kinship arrangements? Do different types of children (or caregivers) become located in formal versus informal kinship settings? Do children move between these two kinship care types, or do children tend to track into one or the other?
Empirical investigation of the use of formal and informal kinship arrangements has been inconclusive, largely due to serious constraints in the data available for analysis. As discussed earlier, there are multiple national data sources that provide information about the prevalence, distribution, and characteristics of children living in kinship care situations. The Current Population Survey, for example, provides detailed estimates of the population of children living with relatives as well as estimates for a population of children defined as "foster children." But, because there is no way for a child to be simultaneously identified as a relative and as a foster child, children in kinship foster must necessarily be lumped into one of these broader categories, either as a relative or a foster child. Our presumption is that most kinship foster cases are defined in the CPS by their "kinship" status instead of by their "foster care" status, so we would expect that the foster care category is comprised mostly of those children living in non-relative foster family placements. This leaves us with no representative national data source that discriminates kinship care cases between "formal" and "informal", and with no national data sources that will allow us to discriminate between "relative" and "non-relative" foster care.
The key population that must be enumerated, then, in order for these comparisons to be made is the "kinship foster care" group (also called "formal" kinship). Once this group is identified, it can readily be compared to the overall kinship care population to provide (by simple subtraction) a means of separating the total kinship care population into informal and formal subgroups. Similarly, the kinship foster care group can easily be contrasted to the total population of children in foster care to differentiate the kinship and non-relative foster care subgroups.
Although this information is not available from any known national data source, it can be obtained for four of the states that report to the Multistate Foster Care Data Archive Project, managed by the Chapin Hall Center for Children. Based on comprehensive individual-level tracking records of children in foster care, the Archive currently can identify kinship foster care cases in California, Illinois, Missouri, and New York. Some personal and case characteristics are available to describe each child in these formal kinship placements, and the use of kinship care placements can be evaluated in context of its role in a child's complete foster care history. It is worth noting that this information is not available for every state that participates in the Archive. Michigan, for example, uses kinship care fairly extensively, but these cases cannot be recognized through their data system in the cases where the placement is arranged by an independent provider agency. Formal kinship care in Texas, in the sense of paid foster care placements, is fairly rare and not specifically flagged in the tracking system. What is unusual about kinship care practices in the Texas child welfare system is a widespread reliance on "semi-formal" kinship care arrangements. As in other states, a child is frequently placed in the care of a relative while the child legally remains in state conservatorship (custody). Unlike most other states, the relative caretaker typically receives no foster care payments or support. Because this arrangement so clouds the line between formal and informal care, and because the few paid kinship placements known to exist cannot be identified in the data records, Texas was not included in the following analysis.
Kinship Foster Care in Four States
The growth in kinship foster care has been one of the more closely watched trends, and hotly discussed topics, in child welfare over the past decade. Where available, the numbers verify that rapid changes in kinship caregiving have indeed occurred. Figure 3.1 portrays recent foster care caseload growth in California, Illinois, Missouri, and New York, and breaks this growth into kinship care and non-related placement components. While the patterns for the individual states differ in interesting ways, the important role of kinship care in foster care caseloads is apparent. In New York and Illinois, kinship placements were clearly the "growth sector" of foster care, either leading or absorbing (depending on interpretation) most of the rapid growth that occurred in each system during the observed period. In California and Illinois, kinship care either almost equals or exceeds other forms of foster care in frequency. In all four states, kinship care has grown at a more rapid pace than other types of foster care. However, none of these states showed a decrease in non-relative foster care cases during the period of growth in kinship foster care, implying that there is no apparent process of simple movement of foster care cases between classifications.(12) Only two observations can be made from these four graphs that suggest the growth of kinship placements might soon approach some limit. First, although it has increased in recent years, the level of kinship care has remained much lower in Missouri than in the other three states. Second, New York State has actually seen a reduction in the size of both components of its foster care caseload from 1991 through 1994.
In most of the analysis that follows, kinship foster care will be addressed at its April 1990 levels in order to allow direct comparison to the enumerated counts from the 1990 census. This is necessary because the census is the only stable and reliable source of information on the comparison population of interest, informal kinship care, across places. When discussing a clearly dynamic phenomenon, analysis based on examination of a single cross-section potentially involves some loss of information. Because our real interest is in the present (1996) and future, the question is whether analysis of 1990 patterns can tell us anything about current relationships. Looking at the four graphs in Figure 3.1 we can see that, in terms of the overall relation of kinship foster care to non-relative foster care, the 1990 levels are similar to the post-1990 levels in each state except Illinois. Although there is no assurance that other attributes of these groups have not changed, their overall levels have maintained the same basic relation across the 5 year interval. In Illinois, the kinship foster care population grew by over 150 percent between 1990 and 1994, so some additional information will be necessary to allow us to consider how the 1990 findings developed here might be relevant to current issues.
State Formal and Informal Kinship Care Populations
By adding new information obtained directly from the foster care case records held in the Archive, the census-based living arrangement categories described previously in the analysis of living arrangement patterns in fifty states can be extended in several ways. Most important for this work, the counts of children in kinship foster care for April 1990 can be subtracted from the count of children living with relatives that was enumerated in the 1990 census to derive a new count of informal kinship cases. Thus, the "related child" group described in the previous section can now be divided into two subgroups: a formal kinship group -- those children observed in kinship foster care through the Archive data; and the residual informal kinship group -- those children living in relative settings who are not observed in kinship foster care. In a similar fashion, we have broken the "unrelated child" category into two parts -- non-relative foster care and other unrelated children.(13) These operations can be performed for any geographic or substantive subgroup for which the census and Archive both track data. For this analysis, we have tabulated informal versus formal kinship for four states, the counties within those states, and for the 0-5 and 6-17 year age groups within each of these geographic areas.
Table 3.1 presents these modified living arrangement data for each the four study states. These numbers are fundamentally the same as the state numbers in Table 2.2a, except that the detail within the relative and non-relative categories is expanded here using the new information extracted from the foster care data systems in each state. Across these four states, which together include over 16 million children (or over one-quarter of the U.S. child population), we observe that almost 400,000 children lived in kinship care settings in 1990, with the preponderance (331,521) in informal kinship care. A substantial, but much smaller, number (61,023) lived in formal kinship foster care placements. The same basic relationship describes the children living in unrelated care situations, where formal non-relative foster care represents only a modest share of the total children who were housed and cared for by unrelated individuals. Thus, while child caretaking through non-parental living arrangements is a relatively uncommon phenomenon (over 95 percent of all children in these states live with one or more parent), the number of children living in each of the four "atypical" care arrangements described here is still substantial.
Living arrangement patterns can be compared across these four states with percentage distributions. Panel B of Table 3.1 shows that the allocation of children across the non-traditional living arrangements varies in the four states being examined. Missouri had very low levels of formal kinship care in 1990 ( 0.5 children per thousand), while New York had levels over ten times as high ( 5.0 children per thousand). In each of the four states, informal kinship care was a more probable care arrangement than formal kinship care. Missouri, Illinois, and New York showed similar levels of informal kinship care (between 17 and 19 children per thousand), while California had a slightly higher level higher level (23 per thousand). The basic state patterns observed for kinship care are once again reflected in unrelated care arrangements, with New York having the higher prevalence of formal child welfare (foster care) arrangements with non-relatives and California having the highest level of non-relative placements arranged outside of the formal foster care system.
Age and Statewide Levels of Care Arrangements
All comparisons of formal versus informal kinship care must be based on information that is available both through the census data and the Multistate Foster Care Archive, because the informal kinship population can be observed empirically only by joining these two data sources. Although a number of characteristics describe the children in kinship foster care, the only personal characteristic that is available to us from the census data to describe the children in kinship living arrangements is their age, defined within two broad categories. Tables 3.2a and 3.2b present the same four-state population of children described in the previous table, now divided into two subgroups -- children 0-5 years of age and children 6-17 years of age.
The overall structure of child living arrangements is similar across states and between age groups. In each of the four states, the preponderance of children live with one or two parents -- the combined four-state percentage of children living with one or more parents is 96.3 percent for ages 0-5 and 94.5 percent for ages 6-17. Older children are slightly more likely to live in a mother-only situation than are younger children, while younger children are slightly more likely to live with two parents or in a father-only arrangement than are the older children.
More marked differences between age categories begin to appear when we look at the distribution of children living in kinship and unrelated settings. Overall, formal foster care placements are used less frequently than informal (or other) care arrangements in both relative and non-relative settings. There is higher prevalence of formal arrangements for children in the younger age group. For the four states combined, 0.49 percent of 0-5 year olds were in formal kinship foster care as opposed to 0.31 percent of 6-17 year olds. Similarly, there is higher prevalence of informal arrangements for children in the older age group. For the four states combined, 2.5 percent of the 6-17 year olds were in informal kinship arrangements as opposed to only 1.2 percent of 0-5 year olds. Figure 3.2 shows prevalence rates (per 1,000 children) by age group for formal and informal kinship care. In each of the four states, informal kinship arrangements for children 6-17 are approximately twice as prevalent for children ages 0-5. In contrast, the prevalence of formal kinship care arrangements is greater for the 0-5 age group in all states except Missouri (where both levels are very low). Clearly, this represents a patterned response to children's care needs in which the youngest children are more likely to be placed in kinship arrangements under the auspices of the formal child welfare system.
Figure 3.2 demonstrates another attribute of kinship care in these four states. Although the fundamental relationship between the children is age, and kinship care type is persistent, this chart also portrays clearly what does and doesn't vary across states. The higher prevalence of informal kinship care in both age categories is markedly constant for these four states. For 0-5 year-olds, it varies only between 10 per thousand in New York to 13 per thousand in California; and for 6-17 year olds between 20 per thousand in Illinois to 29 per thousand in California. In contrast, while the basic age relationships within the formal kinship foster care category are maintained across each state (except Missouri), the prevalence levels for formal kinship care vary widely. The formal kinship care prevalence in New York was over ten times as great as in Missouri, twice as great as in Illinois, and over one-third as large as in California.
The absolute levels and age composition of children in kinship care arrangements observed here across these four states suggests, in the absence of other contextual information, that the many social forces, pressures, and trends that resulted in children living in these alternative care arrangements have acted similarly in these four parts of the United States. However, we also see that these cross-state similarities do not fully hold for the subset of children in formal kinship foster care. Although the basic age relationship observed for formal kinship care tends to be constant across states -- the relative size of the formal kinship population varies between states more than the informal kinship population does. A simple explanation would be that while overall kinship levels result from general social processes and trends that affect children similarly in each state, the specific response of establishing and supporting formal kinship foster care arrangements is highly dependent on local child welfare policy and practice considerations, which can vary across states. This topic will be addressed again as new information is explored.
We have observed that older children are substantially more likely than younger children to live in informal kinship arrangements, that younger children are somewhat more likely than older children to be in formal kinship arrangements, and that each age group is more likely to become engaged in informal rather than formal kinship care. It should be noted that these two population groups do not equally divide the total child population -- indeed, the 0-5 year group contains only just over one-third (36 percent) of the child population. Although the comparative prevalence rates discussed above are analytically most instructive, they do not directly address the actual population impact of these processes. Figure 3.3 presents the actual distributions of kinship care by age and type for all children living in any kinship situation. Because the 6-17 year age group is most likely to live in an informal kinship setting, and is almost twice the size of the 0-5 group, we can see that the net results are numerically dominated by older children in informal arrangements.
Within-State Patterns: Counties and Regions
In the same way that census and administrative foster care data can be combined to allow examination of child living arrangements separately in four states, data from these sources can produce similar information for smaller geographically defined places within each of these states. As was the case with decomposition of formal and informal kinship along substantive lines (e.g. age groupings above), disaggregation of these populations requires that the same criteria be available for both census and the Archive data. The census data can be mapped to many different local-area levels, but the geographic information currently available in the four-state foster care data is organized at the county level. Thus, while census data limited substantive decomposition to two age groups, the Archive data provided by the state agencies limits the geographic decomposition to counties.
The initial plan for this analysis was to systematically compare the formal and informal kinship care levels for counties within each state to search for patterns in their variability that could help us gain insights into cross-state regularities and within-state patterns. Other census-based areal indicators of such factors as ethnic distribution, poverty levels, employment, etc., were to be employed in this investigation. The living arrangement information for counties in each of these four states is presented as Table 3.3. Within each state, the counties are ordered by child population, from largest to smallest. Because some of the counties have very small populations, care must be taken in interpreting the rates and percentages for these places.
Initial inspection of county kinship care patterns revealed one overarching finding: not only are formal and informal kinship care distributed unevenly across places within these four states, as we might have expected, but the practice of kinship foster care is almost exclusively limited to the primary urban areas in most of these states.(14) This is most clearly the case in New York, where over 95 percent of all kinship foster care placements involve children from New York City. In Illinois, the use of kinship foster care is also highly localized, with 69 percent of the state's formal kinship placements in Chicago. Although Cook County, including Chicago, had over 6,800 kinship foster cases in 1990, the county with the next-largest frequency was St. Clair County (East St. Louis, IL) with only 213 cases.
Of the four states observed, the only one showing significant levels of kinship foster care in areas away from the primary urban place is California. Los Angeles County, containing 30 percent of the state's children, generates almost half (47 percent) of the kinship foster care placements. But, formal kinship care remains fairly common in a number of the other larger urban counties -- San Diego, Sacramento, Alameda (Oakland), and others. The prevalence of kinship foster care in this group of counties ranges from about one-half to two-thirds of the 6.1 per thousand level observed in Los Angeles. San Francisco County has the highest rate of kinship care prevalence in the state--at 11.8 children per thousand, it is almost double the rate in Los Angeles.
Although the lack of variance in formal kinship care precluded a full ecological analysis of counties for the four states, we have attached a few county-base indicators, including race/ethnicity, which would have been used for such an analysis, in Appendix 3. Most of the California counties with higher kinship foster care prevalence also have the largest percentage of African American children (Alameda, San Francisco, Contra Costa, Sacramento, and Los Angeles), and African American children are indeed over-represented in their formal kinship caseloads. Only one California county with a high African American population, Solano County, has very low kinship foster care rates. California is also characterized by a large Hispanic population. The levels of informal kinship care tend to be higher in counties with larger proportions of Hispanic persons-- such as Imperial, Tulare, Fresno and Los Angeles Counties.
Outside of California, county-based analysis of formal versus informal kinship patterns is not instructive because almost all of the areal variation in formal kinship is explained by location in the primary urban place. Instead of forcing an implausible method on these data, the analysis has been simplified to examine the differences observed between these primary urban places and the remainder of each state. Summary state totals and subtotals for "primary urban place" and "balance of state" are presented at the bottom of each state subtable in Table 3.3.
Figure 3.4 presents age and care-type prevalence rates separately for the primary urban place and the "balance"of each state. The relationships shown in these graphs confirm the earlier finding that formal kinship care levels are either extremely small or virtually nonexistent in the balance of each state. Even in Missouri, a state with very low levels of kinship foster care, this population is concentrated in St. Louis. Only California shows any substantial amount of formal foster care outside of the primary urban place, and levels for the rest of the state are still much lower than those observed for Los Angeles County. A second observation is that levels of informal kinship care are consistently higher in the primary urban places than in the balance of each state. A final observation can be made by examining only the upper chart in Figure 3.4 -- the graph for the primary urban counties-- as these are the only places where both formal and informal kinship arrangements occur with any regularity. In all four large cities and within each age group, there is apparently a strong inverse relation between the levels of informal and formal kinship care. That is, when informal kinship is relatively high, formal kinship is relatively low, and vice versa. This is most apparent in the 0-5 year old age group. For example, St. Louis, Missouri shows the highest level of informal kinship (24 per thousand) and the lowest levels of formal kinship (2 per thousand) for young children. Conversely, 0-5 year olds in New York City have the lowest observed levels of informal kinship (12 per thousand) along with the highest levels of formal kinship (19 per thousand). Los Angeles and Chicago are each mid-range for both types of kinship care. A similar relationship, though not quite so strong, exists for the 6-17 year olds.
This inverse relation between the prevalence of formal and informal kinship care in the cities has potentially important implications. Although observations from four places do not provide overwhelming evidence, there is clear suggestion here of a possible substitutability relationship between formal and informal kinship care. Overall kinship care rates for both types (formal and informal) combined are remarkably similar across the four cities: for 0-5 year olds they vary between 26 and 31 per thousand; and for 6-17 year olds, they vary between 41 and 46 per thousand. What is different is how these cases have been sorted between informal arrangements and formal foster care placements in different cities. The fact that a formal response is most likely to be invoked in New York City and least likely in St. Louis is objectively clear in these numbers. What is new is the tentative hypothesis that the response, and not the underlying condition, differs between places. This finding suggests that the children being placed in kinship care arrangements are fundamentally similar in these cities, and that differing public actions (due to policy, casework practices, etc.) create the variation in informal versus formal kinship levels. Factors such as local court decisions, agency placement priorities and payment guidelines, and federal reimbursement claiming strategies might be the more productive basis for understanding levels of kinship foster care, per se, than underlying social causes.
Race, Urban Places, and Formal Kinship Care
We have documented differences in the profiles of the support systems created to care for children who end up living in homes where they are cared for by someone other than one of their parents. Although some of these differences are related to the age of the child and the type (formal versus informal) of care provided, the clearest contrasts observed so far have involved the type of place in which the child lives. The primary cities in each state show higher levels of both formal and informal kinship care than other places in these states. We can only hypothesize about the factors underlying this fundamental difference. The largest cities include substantial concentrations of persons in poverty, disproportionate numbers of minorities, and, when considering formal kinship care, huge child welfare agencies and court systems straining in their capacity to handle growing and complex caseloads. Our largest cities have shown many symptoms of social dislocation; problems such as unemployment, drug use, crime and violence, teenaged parenthood, etc. occur in much greater magnitude, if not more frequently, in large urban places.
The national analysis of children in kinship care (Section I) presented evidence of clear racial differences in the likelihood that children will live with relatives other than a parent. (See Table 3.4). Overall, the combined CPS panels for 1989-91 showed 6.2 percent of African American children living in kinship situations, as opposed to 2.4 percent of Hispanic children, 1.2 percent of white children, and 2.1 percent of the children of Asian, Native American, and other backgrounds. However, the data as reported in Section I, do not provide a cross-classification of race and region. Using information from published 1994 CPS results, we have computed the following breakdown for the likelihood of living in all kinship care (formal and informal) for American children ages 0-14 by age, race/ethnicity, metro/non-metro status, and age category.(15)
Within each age-racial/ethnic group, the nonmetropolitan percentages for kinship care either equal the metropolitan percentages or exceed them by up to one-third. Observed age differences are also relatively small, and seem to have an effect only among all Hispanics and metropolitan African Americans. The racial effect that clearly persists, both across age groups and across metropolitan/nonmetropolitan places, with African American levels averaging about five times higher than white levels and about 2.5 times higher than Hispanic levels.
Census tabulations of children living in kinship care unfortunately do not include any racial/ethnic categorization, so the influence of this factor cannot be directly introduced for consideration in comparing formal and informal kinship care for four states. However, ethnic classifications are part of the descriptive information in the records of each child tracked by the Multistate Foster Care Data Archive, and we can identify the race/ethnicity of each child in a formal kinship foster care placement. So, examination of the extremely important racial/ethnic distribution can currently be made only for the formal component of the kinship care population. It is unfortunate that the informal/formal comparison cannot be analyzed fully along ethnic lines. In the remainder of this analysis, we will simplify comparisons by discussing race only in terms of African American and "all other races."
Table 3.5 presents several types of information that help us examine levels of kinship foster care participation by ethnicity/race and by region of the state. The observed levels of kinship foster care for race-region subgroups, measured by prevalence rates per 1,000 children, are presented in the bold box at the center of Table 3.5, and are portrayed graphically in Figure 3.5. It is once again clear that kinship foster care is most common among African American children and in the largest cities. The highest race-region specific rate observed is in Los Angeles, where 28.5 of each 1,000 African American children are in kinship foster placements. The formal kinship prevalence rate among African American children in New York City is the next largest at 21.2 per thousand. In each state, the lowest rates are those for "others" (not African American) in the "balance" of the state. Observed rates for this group (the largest in population size) are as small as 0.2 children per 1,000 in "Upstate" New York. This provides a striking contrast -- the kinship foster care prevalence level for non-African Americans in Upstate New York is less than one one-hundredth of the level observed for African American children in New York City.
In California and Illinois (and probably Missouri), the race effect is the more dominant of these two factors. Within each regional category, an African American child is eight to ten times more likely to be in a kinship foster placement than any other child. In Chicago, where almost one-half the total child population is African American, African American children comprise over 90 percent of the kinship foster care population. As a result, almost two out of every three kinship foster cases in Illinois are African American children from Chicago. It is significant that in both California and Illinois, African American children living in areas outside of the primary city (i.e. in the "balance of the state") are still more likely to live in kinship foster care settings than are the "other" (non African American) children living within the primary city.
New York presents a different pattern. Although both African American ethnicity and a New York City location are each related to higher levels of kinship foster care,the "city" effect appears more dominant. In New York, the city rates for both racial groups are relatively high, while the upstate rates for both racial groups are low. Non African American ("other") children living in New York City are four times more likely to be placed in kinship foster care (8.4 per thousand) than African American children from "upstate" (2.2 per thousand) -- a relation unique among the four states examined.
Because the actual number of children in kinship foster care can be viewed as the net result of a process that applies these race-region specific rates to the race-region specific populations, we must consider the influence of population composition on the final net result. These "marginals," fully independent of any kinship foster care patterns, provide the context within which the differential tendencies suggested by the prevalence rates can operate. The percentage of children in each state that live in the defined "primary urban place" varies from a low in Missouri where St. Louis contains only 7.6 percent of all children in the state, to a high in New York City where almost 40 percent of the state's children live. Racial composition also varies widely between and within states. About one of every eight children in Los Angeles is African American, compared to about one in three for New York City, one in two in Chicago, and almost two in three in St. Louis. African American children are far less likely to live in the remaining portions of any of these states, with the "balance of state" percentages varying from a low of 6.9 percent in California to a high of 9.6 percent in Missouri.
(12) Much more information about kinship dynamics is available from Archive data. Kinship arrangements are likely to be established fairly early in a child's foster care experience. Apart from the earliest short-term temporary custody placements, most children in kinship foster care placements tend to be "pure" kinship cases, and most children in non-relative placements tend to be "pure" non-relative cases. There is not a high level of movement in between the two statuses. Also, kinship cases tend to have a much longer duration than other foster care. See Goerge, Wulczyn, Harden (1993, 1994) for more detailed Archive reporting.
(13) This "other unrelated children" group, which we use here to describe all unrelated children who do not live in foster care, can also be further subdivided. For all children 0-17, it is possible to differentiate "non-relative family foster care," "child welfare placements in congregate care facilities," "other children in institutional settings," "other children in group quarters," and "other unrelated children living in households."
(14) The "primary urban areas" have been defined as follows for the purpose of this work: Los Angeles County in California, the city of Chicago in Illinois, Saint Louis City in Missouri and New York City in New York. These delineations are somewhat arbitrary and they could be quite arguable. However, the concentration of kinship foster care within these areas is so dramatic that issues of precision are rendered moot.
(15) These numbers are based on computations from document P20-484,Table 3. It should be noted that the "Metropolitan" category is much more broadly defined than our construct of "primary urban place", and includes much smaller cities and suburban counties. Also, children 15-17 are not included in this tabulation..
