PACE WORKING PAM Slamtas 02-3 Child Care Demand and Supply under CaIWORKs: The Early Impacts of Welfare Reform for California's Children, 1998-2000 INSTITUTE OF GOVERNMENTAL STUD???» {TRRARY a 3 2w UNIVERSITY or CALIFORNIA Policy Analysis for California Education University of California, Berkeley October 2002 1. I: fa. K :11 ' «. 3’!» .9 y‘ . , v, , _ ~o *‘fl‘n’x . \, 3‘ WORKING PAPER SERIES 02-3 Child care Demand and Supply under CaIWORKs: The Early Impacts of Welfare Reform For California's Children, 1998—2000 Diane Hirshberg University of California, Berkeley October 2002 A Policy Analysis for California Education PACE Acknowledgments This paper stems from the Child Care Planning Project of the California Department of Social Ser- vices. Special thanks to Lyn Vice, Hilva Chan, Valerie Maulet, Teri Ellen, and LaRae Bustamante (CDSS), and to Nancy Remley at the California Department of Education, for their long-range perspective and steady support. Tom Burke and Paul Smilanick provided much of the data considered in this and other project studies. Additional funding and related research is supported by the Child Care Re- search Consortium, supported by the Child Care Bureau of the US. Department of Health and Human Services. PACE’s involvement in the consortium is made possible by the California Child Care Resource & Referral Network. Special thanks go to Patricia Siegel and Shelley Waters-Boots. No agency necessarily supports the findings or interpretations appearing in this paper. Table Of contents Executive Summary ............................................................................................................................................ 1 Introduction ............................................................................................................................................................. 3 Child Care Demand Issues ................................................................................................................................ 5 Child Care Supply and Growth ...................................................................................................................... 3 The Evolving Policy Context .......................................................................................................................... 4 Study Methods ....................................................................................................................................................... 5 Child Care Demand Issues ................................................................................................................................ 6 Understanding Subsidy Utilization ............................................................................................................ 6 Estimating Subsidy Use for CaIWORKs and Low-income (non-CalWORKs) Families ........................................................................................................................................................... 8 Characteristics of Parents Using Child-Care Subsidies ................................................................ 9 The Kind of Care Chosen by CaIWORKs Participants Utilizing Subsidies ...................... 15 Types of Child Care Selected by CaIWORKs and Low-income (non-CaIWORKs) Parents in the Three Counties .................................................................................................... 16 Child Care Stability ............................................................................................................................................ 19 Child Care Supply Issues ................................................................................................................................ 2O License-Exempt Care ....................................................................................................................................... 23 Disparities in Child-Care Supply and Growth ................................................................................... 24 What community factors predict growth in licensed child-care capicity? ............... 25 Continuing Questions ...................................................................................................................................... 26 Appendix A ............................................................................................................................................................. 28 Endnotes .................................................................................................................................................................. SO .‘ ,p 7.1%?" y.“ vs , w, 4 . ' MW ”1 1 v; > t. A, g a «, 34'0‘§>\-1‘;1x—.L:1.~v«4 . .. ”q. {KNEEI ' ‘ V , ‘I‘ _, V ; g. , ‘ 4 {Ma w: » ‘ ‘ a»), ‘. 9v “ x. ( 1 ’. I» 1 A 4‘21”» 1 »; “r , ‘ V, u ‘ , ~ ‘ m, ' 1 - mm. 5 = ”wwwfimmnuuuwm u,»...,.u.u.y.y mg ’ i$m .4 , . ‘Mt.l.}‘1>-1$fi"v‘:“i§";ffz,nl “hug“.um u n. l I , \ g . _ \ 1,‘ ’95 ,. t _ 3E * v, ' «3.: , .mn‘w.‘ » ~r1v"“i‘l'1f~l( ‘ ”.nmwvx... n.”gape...”n‘xmu..« . ‘,v...u_ ”HM V . . ‘ » .sL ' 1‘”; H , . .“,/;.Mm ,nw.‘ «ugmnvyhfluu.H» ‘ .w . U, .. ”A. . {N .‘ “1 . ‘ _ 45 ‘ ‘ ~ 4 ‘, . ~ ~ ,,,,,, -:-\(VKVV‘QB¢|:-W!>.O$.» ‘ «5» -, v. - I > g? I ' ‘g P 'W‘fiamm W ‘“ w 3 -' ‘ “ ‘ '.-_§ <. H g ' {2: ‘ {g “13“.. 5:} 13‘ tux i 3‘1“"? 4, ' i ' ~ ‘ -) I. ‘6 ‘4 x w 3. , ., .mcmu-‘w ‘~‘,....H u . ‘ . 7“» u . . . . . .V . ‘ ‘v a \ f 4;. a . . .. 1. 4- .1. u- 1. \ Executive Summary In 1996 the federal government passed the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), which included the Temporary Assistance for Needy Families (TANF) block grant. TANF altered the structure of the welfare system nationwide and prodded millions of welfare recipients into jobs or welfare-to—work activities. California‘s reform program, the California Work Opportunity and Responsibility to Kids Act, known as CalWORKs, was enacted in 1997. The TANF and CalWORKs welfare reform initiatives included significant changes in the work support systems for parents on public aid. In California this led to both a restructuring of the child-care subsidy system, as well as new investments in child—care capacity building efforts. Prompted by changes in the child-care system and welfare reform, the California Department of Social Services (CDSS) asked PACE researchers to look at how new welfare-to-work and child-care capacity building were affecting supply and demand in the child-care system. This paper provides an overview of our findings in response to this request. Subsidy use a The number of families using stage 1 child-care subsidies increased in 1999 and early 2000, even as CalWORKs caseloads declined. However, the number gradually decreased through the remainder of 2000. It is difficult to interpret the meaning of these changes, as we cannot obtain data allowing a comparison of these trends to changes over time in the use of stage 2 child—care subsidies. a Calculating subsidy utilization rates is difficult. Determining who uses subsidies and who is eligible for subsi- dies is complicated both by data limitations and by questions regarding whether all eligible parents prefer to use child care subsidies. Child care choice E The type of care used by parents was closely affiliated with ethnicity, language, age of children and whether they were CalWORKs participants or receiving subsidies via the APP (non~CalWORKs) subsidy streams. White parents were more likely to use licensed care than Latino or black parents, and Spanish—speakers used exempt care more often than English—speakers. g Parents with infants were more likely to use family child—care homes, while those with older children were more likely to use center—based care. Rt? Finally, CalWORKs parents were more likely to use exempt care versus parents accessing the APP (non— CalWORKs) child-care subsidies who tended to use center-based programs. Stability in child care , CalWORKs parents reported fairly stable child—care arrangements. Across family samples from in counties, between 60% (Alameda County) and 83% (Los Angeles County) of the parents used the same child care pro- vider for four or more months. The use of the same provider over a four-month period was even higher for APP (non-CalWORKs) clients: 93% in Los Angeles and 85% in Kern County, although over seven months it dropped to about 70%. Child care supply issues Expansion in the supply of licensed child care is barely keeping pace with growth in population of children, age 0-5. It appears that in centers much of the growth may be in newly created slots in existing facilities, rather than resulting from development of new centers. While the supply of care in Family Child Care Homes (FCCHs) has increased, some of it may be due to changes in licensing rules, where existing homes could serve more children, rather than increases in total number of facilities. The growth was unevenly distributed, with more homes being opened in predominately Latino zip codes, than in non—Latino zip codes. This helped to offset some of the disparities in the availability of center- based care in Latino communities. The distribution of licensed centers is not equitable across communities. A greater supply of center-based care is found both in affluent neighborhoods and in communities where the average educational attainment of parents is higher. Factors predicting expansion in licensed care are difficult to identify when examining growth patterns between 1996 and 2000. Growth in FCCH capacity was lower in zip codes with higher maternal employment rates and in zip codes with stronger child—population growth Center capacity for infants. age 0—2. increased between 1996 and 2000, but decreased for 2-5 year olds after taking into account child population growth. Introduction In 1996, the federal government passed the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA). which included the Temporary Assistance for Needy Families (TANF) block grant. TANF altered the structure of the welfare system nationwide and prodded millions of welfare recipients into jobs or welfare—to- work activities. California’s TANF initiative, the California Work Opportunity and Responsibility to Kids Act, known as CalWORKs, was enacted in 1997. The TANF and CalWORKs welfare reform initiatives included significant changes in the work support systems for parents on public aid. The framers of these new policies recognized the importance of providing transportation, child care and other supportive services to enable parents to enter into and succeed in the workplace. In California this led to both a restructuring of the child—care subsidy system, as well as new investments in child-care capacity building efforts. Prompted by changes in the child—care system and welfare reform, the California Department of Social Services (CDSS) asked PACE researchers to look at how new welfare—to—work and capacity building initiatives were impact— ing supply and demand in the child-care system. The policy questions they posed were, broadly, as follows: Child care demand issues E To what extent is demand for child care, both licensed and license—exempt, rising among CalWORKs parents? E As demand rises. is the percent of parents using child—care subsidies increasing? $3 How many families are using Stage 1 subsidies, or other forms of child care subsidies? For how many months do CalWORKs parents typically utilize Stage I or other child—care support? E What types of child care are CalWORKs parents selecting? Do these selection patterns vary among ethnic and language groups or counties and are they changing over time? E Do parents using CalWORKs child-care subsidies select different kinds of care when compared to parents receiving subsidies from the (non—CalWORKs) Alternative Payment Program? E How stable are the child—care arrangements selected by CalWORKs parents? Do they tend to move their children from informal to licensed child—care providers as their work situations stabilize or their youngsters grow older? Child care SUDDIV and growth E Is the supply of licensed child-care programs — centers and family child care homes (FCCHs) — growing at a steady rate to serve an expanding share of low—income families? As more child—care vouchers are funded, does the rise in parents’ purchasing power help to drive the expansion of licensed child care? This paper provides an overview of the findings of the research efforts that were driven by these questions. The evolving DOIICV context Recent Trends. Rising Public Investment in Child Care. In California, more than 1.4 million families, with children under 18 years of age, are living below two hundred percent of the federal poverty level (in the year 2000, the poverty threshold for a family of four was $17,463).‘ And in those families, both in two-parent and single- parent households, there is at least one adult who worked at least part-time during the year.2 In single—parent families and in households where both parents are employed, those who wish to remain employed must rely on child—care institutions or informal arrangements with kith or kin members to provide care for their children while they work. Over the past decade state policymakers have come to see the public funding of child—care and preschool programs as a crucial support if low-income parents are to stay on the job and off the welfare rolls. Over the past four years, Califomia’s annual investment in child—care and preschool programs has almost quadrupled, rising to $3.1 billion in fiscal year 2001-2002, including state and federal dollars. This climbing investment is energized by an evolving public interest in child care for low-income families. Government agencies, increasingly, are also supporting after—school programs. The policy theory in each instance is that by targeting child-care investments on low—income families—beyond the basic values of equity and fairness—all Californians will benefit economically. New challenges to the system. The importance of understanding child care and subsidy use among CalWORKs parents is heightened by the impending federal reauthorization of the Temporary Assistance for Needy Families (TANF) program. In addition, Congress must decide on the reauthorization of the Child Care and Development Block Grant (CCDBG) by the end of 2002. Following these federal actions, California legislators may need to modify the CalWORKs program. In addition to potential changes resulting from the reauthorization of TANF and the CCDBG, California’s child- care subsidy system will be facing other challenges. In 2000, Governor Davis initiated a review of state child—care fiscal policy by the State and Consumer Services Agency, prompted by concerns that while spending on child care has been increasing, the monies are not being used as efficiently as they might be, and subsidies might not be reaching all needy parents in an equitable manner. Approximate 200,000 to 300,000 low—income families are currently on the waiting list for a subsidized child—care slot. In addition, California is facing a dramatic increase in child—care costs for former CalWORKs families. The cost alone to provide subsidized child care for former recipients is increasing from an average of $130 million a year to $747 million in 2005—6. Without such support these families may lose theirjobs, return to welfare, or leave their children in inadequate care. At the same time the economic picture nationally and in California is changing quickly. An economic slowdown is shrinking state revenues and forcing cutbacks in spending. The impact on child—care subsidies is yet to be deter- mined. but most likely there will be less money available to serve a growing population of low—income families who need assistance. And, all of this comes on the heels of a significant change in attitudes about the role of child care in our state and society. 4 A change in thinking. How has welfare reform in California shifted our thinking about the role and utility of child care and preschool? a: First, single mothers with children under 3 years of age were rarely subjected to work requirements prior to the 1996/97 welfare overhaul. These parents are now subject to work activity requirements, such as participation in a job club or other work-preparation activity. In addition, due to funding constraints fewer mothers with children over age 3 had been required to work under the state’s earlier, more modest efforts to move clients from welfare to work. a Second, the time limits on cash aid under CalWORKs —— five years lifetime — mean that single mothers are expected to become economically self—sufficient. However, these time limits have raised questions as to which work supports are needed for women to enter and stay in the workforce, as well as if poor women with limited skills will be able to continue employment if they face sizable child-care costs after they leave cash assistance. fi Third, welfare reform has contributed to the decline of the welfare rolls nationwide, with more modest declines in California. But studies of the families who leave the cash assistance system reveal that many fail to rise above the poverty line.3 This means that the ranks of working—poor families are growing and the welfare—poor are shrinking in number. If our society is serious about encouraging work and upward mobility, then again, the crucial supports and incentives for parents to stay in the labor force must be strong and accessible. While the findings presented here do not address the larger structural issues of welfare-to—work reform, we believe that they do provide insight into the choices and behaviors of CalWORKs parents regarding child—care issues. These we hope will be useful to all child care planners including state and local policymakers, program directors, and others focusing on how to best meet the child—care and early education needs of young children in California. Study methods This paper covers the results of the research done in the early part of the CDSS—PACE Child Care Planning Project. Three counties, selected through a competitive application process, acted as our research partners for this work: Alameda, Los Angeles, and Kern. They were not chosen to be representative of the state of California as a whole. Instead, they offered perspectives from urban northern (Alameda), urban southern (Los Angeles) and rural/ suburban central valley (Kern) communities. In addition, these areas are socially, racially and economically diverse, which allowed us to look at differences across language and ethnic groups. Finally, under the state’s decentralized welfare-to—work system, they each implemented slightly different models for administering CalWORKs child—care subsidies.4 From 1999 to 2000, with our state-level colleagues and local counties, PACE researchers coordinated these core activities: % Client and child-care tracking. Our county partners pulled data on selected CalWORKs and low—income (non- CalWORKs) families related to their child~care subsidies and types of care providers selected. By combining caseload data from welfare departments and child-care Alternative Payment Program (APP) agencies, we were able to compare CalWORKs and APP (non-CalWORKs) families.5 For all three counties we pulled client data at three points in time during calendar year 1999. a Statewide child-care trends. We studied statewide trends in subsidy use and types of care selected by CalWORKs parents between late 1998 (baseline) and late 1999, assisted by the CD58 Research and Evaluation Division, Data Analysis and Publications Branch. as Planning data on the distribution and growth of child-care supply. PACE worked with the California Child Care Resource & Referral Network to obtain 2000 data on the supply of licensed child-care facilities across counties and among zip-code areas within counties. The availability of 2000 data also allowed the opportunity to study growth in the capacity of centers and FCCHs between I996 and 2000. a Focus groups with CalWORKs and APP clients and child-care providers. PACE researchers conducted a series of focus groups in the three counties with CalWORKs participants, low—income working parents accessing APP subsidies. and child—care providers. Reports on these focus groups will be available under separate cover. This paper is organized into two parts. The first looks at the child care demand issues for CalWORKs and working low-income parents, as detailed in the questions posed above. The second addresses supply issues. We finish by raising the questions that emerged from this work. which need continuing investigation. Child care Demand ISSUES In this study, we looked at two central issues related to the demand for child care among CalWORKs parents— subsidy utilization and the choice of a child—care provider. Understanding SlleidV utilization Childncare subsidies are a policy strategy that are intended to assist low—income working parents by raising their purchasing power and thus increasing their child—care options. It is recognized that child care is critical to CalWORKs parents’ ability to enter and succeed in the workforce, but quality child care is often out of the reach of these parents unless they have a subsidy." In California. subsidies are provided in two forms; the first is a voucher which parents can use to pay for care in a center, FCCH, or from a license-exempt individual, related or not. The second is via access to a state or federally funded slot in a child—care center or preschool. In our investigation of subsidy utilization, we looked at trends in the use of Stage 1 subsidies across the state, and then delved into data on those parents accessing subsidies on the local level. Statewide trends in utilization of subsidies. Between the last quarter of 1998 and the first quarter of 2000, there was a steady increase in the overall number of families using CalWORKs Stage I child-care subsidies. from just over 45,000 to over 54,000. This occurred despite an overall decline in the CalWORKs caseload over the same time period. While it appears that the proportion of eligible parents using subsidies was increasing, it is not possible to draw a definitive conclusion without equivalent figures for Stage 2 subsidy utilization by parents still enrolled in CalWORKs and on cash aid. One of the problems in understanding statewide and county utilization of subsidies is that due to administrative and budgetary constraints. a significant number of CalWORKs clients that were defined as Stage 2 participants were covered out of Stage 1 funding for a period of time. Thus it’s hard to know whether some of the increase in the number of CalWORKs clients utilizing Stage 1 subsidies was due to absolute increases in the number taking up their subsidies or the addition of new CalWORKs clients to a population that should have been in Stage 2 programs. Another challenging factor is that Stage 2 clients may still be receiving cash aid, or may already be off of CalWORKs cash aid and into the 24 month period during which they are entitled to child-care subsidies. This means it is impos— sible to determine precisely the percent of current CalWORKs clients that are using child-care subsidies. If there was a proportional decline in Stage 2 rolls, then there was no overall gain in the proportion of CalWORKs participants using subsidies. Conversely, if there is an increase in the number of Stage 2 clients receiv- ing cash aid and using subsidies. then the proportion of clients using subsidies may have gone up substantially. The number of families using Stage 1 subsidies declined during 2000. This decline was matched in the second half of the year by a decline in the number of adults participating in CalWORKs activities. FIGURE 1 Count of Families Receiving Stage 1 Subsidies Statewide as Percent oanVaseline OUarter (Oct-Dec. 1998)7 30.0% — 25.0% /’\ 20.0% / \\ 15.0% ‘3.) 5 / \ g 10.0% m D- / 5.0% / 0.0% v -5.0% -10.0% , . . 1 . t ‘r 1 Baseline: Q1:Jan.- Q2: Apr.- Q3: July— Q4: Oct.— QS:Jan.— Q6: Apr.— Q7: July- Q8:Oct.- Oct.—Dec. Mar. June Sept. Dec. Mar. June Sept. Dec. 1998 1999 1999 1999 1999 2000 2000 2000 2000 Quarter Estimating subsidy use for caIWORKs and low-income (non-CalWORKs) families Calculating child-care subsidy utilization rates can be difficult, as there are several challenges involved. The first is simply determining who is and is not using subsidies. There is no comprehensive list of parents who utilize all sources of subsidized child care within counties or across the state and data systems at different agencies are not linked. While CalWORKs program administrators know how many Stage 1 and 2 subsidies are being provided to program participants, in those counties where Stage I and 2 care are contracted out to APP agencies who cannot enter data back into the county social services databases (e.g., Alameda County), they may not know which persons specifically are receiving them. In addition, parents may be receiving more than one subsidy if they are using multiple care providers or programs. Another problem comes in determining who is really eligible for subsidies—the denominator of this equation. This issue is complex both when looking at families on CalWORKs and low—income working families in general. For example. although a family may be low-income, they may not qualify for a child-care subsidy because they are not engaged in work or education activities that make them eligible. This is particularly important when trying to determine how many children in families participating in CalWORKs are eligible for subsidies. In addition. there is the question of whether all of the eligible parents need or want a subsidy. Some families will not want subsidies because there are older siblings or other relatives in the household who will provide care without cost. It is difficult to know whether families make these choices due to preferences, or because they are unaware of the subsidies for which they qualify. In our investigation of subsidy use on a local and statewide level, we ran into multiple factors that prevented the determination of the exact number of CalWORKs parents using child care subsidies. In only one of the three counties. Kern. are child-care subsidies administered entirely by one agency with one database accounting for all recipients. But in Alameda County the CalWORKs and APP (non—CalWORKs) subsidies are administered by multiple agencies, and they do not have common databases or information. This results in the CalWORKs program having a count of participants receiving CalWORKs child—care subsidies. but not knowing which participants are or are not using the subsidies, and which might be accessing funds via another source, as is described earlier in this section. An effort by county administrators to match this data by hand resulted in the identification of a number of children in Head Start programs that were also in families on CalWORKs. Trying to calculate the subsidy take—up rate among CalWORKs clients in Alameda County using only information from the CalWORKs program thus may lead to an underestimate. assuming that there are a number of families accessing these other programs. Within our three research counties. we attempted to find a way to understand and calculate subsidy take-up rates among CalWORKs and low—income (non-CalWORKs) families. We gathered data on the number of children receiving child—care subsidies via CalWORKs. the Alternative Payment Program. Head Start. and state—funded preschool slots in each county for nine to ten zip codes that had a high number of CalWORKs participants. In order to calculate the percent of the eligible population utilizing subsidies. we estimated the number of children age 0—5 in families whose incomes fall under 75% of the state median income (SMI). and compared this with the number of 8 children receiving known child-care subsidies“. The resulting estimates of the subsidies being used for eligible children ranged from a low of 11% in one zip code in Los Angeles County to a high of 86% in Alameda County. And in all three counties the range in utilization rates among the zip codes was considerable. However, there were some holes in the data we used for our estimates that made it imprecise at best. Perhaps most important, our data did not allow us to account for families who reside in one zip code, but utilize subsidized child-care services within another zip code, as this would require data that we did not have. Our esti- mates were not able to differentiate between children whose parents are engaged in subsidy-eligible CalWORKs activities versus those whose parents are not. Moreover, we were not able to factor in the percentage of one—parent families and those with two parents where both are working versus two—parent families where one may stay. at home. In 1999, Kern County planners did try to look more closely at CalWORKs subsidy utilization using similar methods. They calculated subsidy use estimates for selected zip codes based on the number of children in non— exempt CalWORKs households and the number of children receiving a CalWORKs child care subsidy. The CalWORKs subsidy use figures ranged from a low of 12% to a high of 37%. These figures did not, however, include use of state—funded preschool slots or Head Start programs. Finally, in trying to understand who is using subsidies, it is important not to lose sight of the question of who wants to use subsidies. We cannot assume that all eligible parents want subsidies, nor can we assume that parents who are not using them have chosen not to access them. This means that policymakers need to build some flexibility into the subsidy financing system. to allow for changes in parent choice and demand. Characteristics of parents using child-care subsidies While we could not calculate the number of CalWORKs clients using all forms of child care subsidies across the state or within counties, we were able to learn about some of the participants using subsidies in our study counties. In this section we take a look at the characteristics of parents using subsidies via both CalWORKs and the Alterna— tive Payment Program (APP) (non—CalWORKs) across Alameda, Kern, and Los Angeles counties. What we know about who utilizes subsidies. Subsidy utilization rates vary widely across communities and populations. Although it can be difficult to determine precisely how many CalWORKs parents and low—income (non—CalWORKs) parents are using subsidies, earlier work by PACE researchers has shown that who do not use subsidies are typically: % Parents from immigrant communities, including Latina and Vietnamese mothers. E Parents with children under 3 years old who believe that subsidies are fused to center—based programs, unaware of the options available with child—care vouchers. a: Parents with stronger support networks who often find a kin member to provide child care, losing out on voucher support. @ Parents with no prior experience with welfare or center—based child care, those with the least knowledge of subsidy options. m Parents who live in lower middle-income neighborhoods with a scarcity of centers and family child-care homes.9 In Alameda and Los Angeles counties, we looked at a sample of clients to examine the demographic factors that appear related to the subsidy choices that parents make and how families receiving child-care subsidies via CalWORKs compare to those accessing the APP (non-CalWORKs) subsidy system. In Kern County we were able to examine all of the parents utilizing subsidies via CalWORKs and APP funding streams. However, due to different program reporting requirements, not all of the same demographic information was available for all of the clients studied here. Kern County. As we noted earlier, in each county we collected data at three points in time. In Kern County, these were March, June and September 1999. Ethnicity data was available only for CalWORKs participants receiv— ing Stage 1 and 2 subsidies, not for APP low—income (non-CalWORKs) parents. The CalWORKs population in Kern was primarily white, Latino and black. We found marked differences between the population that was using subsidies and those who were not using subsidies. While the population using subsidies is 43% white, 34% Latino, and 22% black, the population not using subsidies is 47% Latino. 37% white, and only 14% black, as is shown in Figure 2. The reasons behind this are unclear. and there are many possible explanations that need to be investigated further.” FIGURE 2 Ethnic Breakdown of CalWORKs Child-Care Subsidy Users and Non-Users, Kern County March 1999 50.0% - 40.0% 'L'“ 30.0% +— ‘ 1:1 subsidy users 20.0% “‘- non-subsidy users 10.0% ~— 0.0% 1 White Hispanic Black Other Differences in primary language between subsidy users and non—users in Kern reflect a similar pattern (see Figure 3). Language use among Kern county subsidy recipients is almost exclusively English. While 99% of subsidy recipients are English speaking and 1% Spanish speaking for the month of March, 18% of subsidy non— users are Spanish speaking and 82% speak English. Other language groups make up less than one percent of these populations. 10 FIGURE 5 Language Breakdown March 1999 100.0% 80.0% of CalWORKs Child-Care Subsidy Users and-'Non-Users Kern County, 60 0% 40.0% [1 subsidy users nonsubsidy users 20.0% 0.0% Spanish English Other The ages of Children in the families utilizing subsidies differed between those in the CalWORKs and those in the APP subsidy programs. The children of parents receiving APP (non—CalWORKs) subsidies have proportionately more preschool and school—age Children. while toddlers and younger Children are more numerous among parents in the CalWORKs program. FIGURE 43 Age Distribution of Children, Kern County CalWORKs families, Stage 5 1 & 2 June 1999 Frequency 2345678910111213 Age FIGURE 4b Age Distribution of Children, Kern County APP (non-CalWORKs) families, June 1999 Frequency i Age 8 91011121314 123456 11 Alameda County. In Alameda County. data on a small sample of CalWORKs Stage 1 clients (306) were col- lected for the months of June and September 1999. and January 2000." In addition, a one—time sample of I75 Stage 2 and 3 clients was obtained from one large APP agency, Bananas, in February 1999. In all of the samples, African- Americans constituted the majority of all clients. English was spoken by more than 90%. A majority of the Stage 1 clients using child—care subsidies in Alameda (see Figure 5) were engaged in employ- ment activities. with the next largest group participating in training programs. FIGURE 5. CaIWORKs Activities of Alameda County Stage 1 Clients, June 1999 Sample 2% \ I Working i I Cal-Learn ; CaiWORKs ‘ County Respite ‘ Training Training/ Working D i [:1 Job Search i a I l] Los Angeles County. In Los Angeles County. data were collected from five APP agencies in the months of September and December of 1999. Ethnicity information was available for 3.004 of the 5,527 CalWORKs clients. A1 54.4%. Latinos represent the largest ethnic group receiving child—care subsidies. followed by white at 11.1%. Of the 3.485 families receiving subsidies through the APP program, ethnic data was provided for 1.326 children. There were 720 Latino families receiving subsidies through the county‘s APP program (54.3% valid percent), followed by 129 White families (9.7% valid percent). Similarly. for the month of December, Latinos were the majority of the clients for both CalWORKs and APP (non—CalWORKs) subsidy users. Data on Stage 1 clients in Los Angeles was collected in February. May. and September 1999. The population of (‘alWORKs clients utilizing Stage 1 subsidies looked very different from clients accessing Stage 2 CalWORKs and APP (non—CalW(,)RKs) subsidies. The proportion ofAfrican American clients was higher than in the other populations. The language population in Los Angeles consisted of more than a dozen different languages. The most common language spoken by CalWORKs Stage I and 2 participants as well as APP (non—CalWORKs) clients was English. However. the percentage of Spanish speakers was comparatively higher than in Alameda or Kern County at all three points when data was collected. The highest percentage of Spanish speakers was among APP clients for the months 12 of September and December of 1999, at over 20%. In addition, there were fewer English speakers among the APP clientele and the remaining dozen languages made up a negligible percentage of the total. As in Kern, the youngest children in Los Angeles CaIWORKs Stage 1 families tended to be younger than the children in the APP (non-CaIWORKs) families. However, Stage 2 families tended to have older children than Stage 1. In Los Angeles, the majority of clients receiving Stage 1 child—care subsidies in May 1999 were participating in education programs, known as “Self—Initiated Programs.” The CaIWORKs component data for September and February 1999 indicate similar proportions. As was noted before, in Los Angeles clients are considered “stable" and eligible for Stage 2 child—care subsidies when child care and employment or training are “stable.” Generally, em- ployed clients are considered Stage 2 clients, while those participating in the components shown below in Figure 8 are considered Stage 1. FIGURE 6. CaIWORKs Component, Los Angeles Stage 1 Clients, May 1999 1 % unknown domestic violence services job club mental health services substance abuse services self-initiated program DUDE!!- 1 °/o The kind of care chosen by CaIWORKs participants utilizing subsidies statewide A concern of many educators and policymakers is whether low—income parents are able to access quality child care. Subsidies allow low-income parents to afford child—care options that they might not have had previously, including access to high—quality child—care centers which otherwise would be beyond their financial reach. In this section we examine the utilization of licensed care among CaIWORKs parents to see if low—income parents with subsidies are indeed selecting licensed care. The utilization of licensed care among CaIWORKs Stage 1 participants varied widely. The statewide average at of the end of 1998, our “baseline” period, was about 38%. However, in Los Angeles only 20% of these clients were accessing licensed care, while in Alameda County almost 50% utilized licensed care (see Figure 7). 13 FIGURE 7 Percentage of Stage 1 Children in Licensed Care, Oct. —' Dec. 199812 I 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% There are a number of factors impacting parents’ selection of licensed or license-exempt care”. For example, in Los Angeles, Stage 1 clients are typically in job search or education programs, and get transferred to Stage 2 when they find stable employment. Thus, they may not be as stable as Stage 1 subsidy recipients in counties that do not transfer working clients to Stage 2 until they are off of cash aid. Clients who are in unstable or temporary situations are less likely to use licensed care than those with steady employment. In addition, prior research has shown that low-income communities, in which many CalWORKs parents reside, often have less access to licensed care than more affluent communities.14 Furthermore, the overall supply of licensed care within counties may also impact parent choices. The graph below (Figure 8) shows that among the counties displayed, Los Angeles County—in addition to having the lowest utilization of licensed care among its Stage 1 population—also has the lowest supply of licensed care per capita. Conversely, Alameda County has both the highest proportion of licensed care to the population as well as the highest utilization of licensed care among its population. FIGURE 8 Supply of licensed child care per 100 children age 0 — 5 20 18 aw_ 16 — Q _ 14 ~— __ 12 —— M 10 _ __ 8 __ __ _ 6 “—l ».~__ ___ __ _ ._ 4 __ __ _ ~— 2 __ i.— .___ 0 l l I I I l I l I D t -d d b th a a orovr e y e 442% (90% ($86}? «fig/'0 {out 450% £50 450$ £047, «53% California Child Care 6% 0 (7060 90 ’60 ’50 Q ‘50, Resource & Referral G/GJ, [2’0 (30 0600 “9% 6 Network, 1998. 14 While the availability of licensed care isn’t the only factor explaining parents’ choice of licensed versus exempt care. it can be a significant influence. When we examined the use of licensed care by Stage I CalWORKs participants over the course of a year, we saw a trend toward a higher proportion of parents statewide selecting licensed care. but also a lot of fluctuation within the counties (see Figure 9). The most dramatic shift occurred in Los Angeles County, where the proportion of Stage 1 subsidized children in licensed care increased steadily over the course of a year. One reason might be that in Los Angeles, the CalWORKs program was fully implemented later than in other counties, and co—location of resource and referral agencies may have taken longer. Moreover it took time for the contracts with APP agencies for Stage 1 services to be signed and enacted. Finally. as more clients in Stage I became stable, they were not trans— ferred to Stage 2 services due to budget constraints. Thus the numbers could represent some of the Stage 2 clients, who as participants in stable work or education situations are expected to use licensed care more than those in Stage 1. Without more systematic data on how these parents selected their care, it is hard to explain this change fully. FIGURE 9 Percentage of Stage 1 Children in Licensed Care15 Oct. 1998 - Dec. 1999* 60.0% + Alameda *5” Butte »/~: Fresno Kern Los Angeles + Sacramento ' San Diego + San Francisco " , t L: Santa Clara 20.0% 1 fl 1 1 Baseline: Oct.— Ql : Jan-Mar. Q2: May-June Q3: July—Sept. Q4 Oct. - Dec. . Dec. 1998 1999 1999 1999 1999 “‘— Stats/Wide 15 Types of child care selected by CalWORKs and low-income (non-CaIWORKs) parents in the three-county sample. In the first part of this paper, we talked about who was accessing child—care subsidies in our three-county client tracking sample. But our study also looked at what kind of care parents are choosing, and how the choice of care dil‘l‘ers among parents according to a host of factors. We examined the use of CalWORKs Child-care subsidies versus APP subsidies. as well as ethnicity. language, and the age of the child. Kern County. Parents using CalWORKs Stage I and 2 subsidies selected license—exempt care at a greater rate than parents using other APP child—care subsidies. In March 1999, more than 50% of the APP clients selected center—based care compared to an average ol‘ 2] % ol' the CalWORKs clients using subsidies across all panels. In June. the breakdown of care used by Stage I and Stage ’2 subsidy recipients was almost identical: 58% of Stage I recipients used exempt care. compared to 57% of Stage 2 recipients, and l9% of stage 1 recipients used centers compared to 23% ol‘ stage 3. FIGURE 10a Type of Care, CaIWORKs Stage 1 and 2 clients, Kern County, Sept. 1999 Exempt I Individual [:1 Center D FCCH FIGURE 10b Type of Care, APP (non-CaIWORKs) clients, Kern County, Sept. 1999 Exempt I Individual [:1 Center [:1 FCCH 16 The type of care selected varies by both ethnicity and language for CalWORKs clients but was fairly consistent. White and English-speaking clients were more likely than other groups to choose center-based care. Kern county clients also varied their choice of care depending on the age of their child or children. Among APP families, those with infants were more likely to choose family child—care centers, while those with toddlers and school—age children were more likely to select center-based programs. Finally. our analysis looked at the type of care selected in Bakersfield compared to those in rural Kern County. first by CalWORKs Stage I or 2 clients and then by other APP (non—CalWORKs) subsidy recipients. In rural areas. all clients. regardless of what type of subsidy they were receiving, were more likely than those in urban areas to use license-exempt care. In Kern County we also looked at whether the type of care chosen was related to parents using care within their home zip code and found that CalWORKs clients were slightly more likely to use care in the same zip code where they live than APP subsidy recipients. Alameda County. As noted earlier, the data collected for Alameda County consists of samples drawn from a larger sample of 449 clients selected in March 1999 from the files of the two APP agencies administering Stage I subsidies in the county. Community Connection for Child Care (4Cs) and Child Care Links, and a file of clients from one of the largest Stage 2 providers. Bananas, drawn in February 1999. Slightly over half of the children in the sample of Stage 1 clients were in licensed care, compared to slightly less than half of the Stage 2 clients. (see Figures lla and llb). FIGURE 11a Type of Care Used by Stage 1 Clients, Alameda County June 1999 Exempt I Individual [:1 Center E] FCCH 17 =FIGURE 11b Type of Care used by sample of Stage 2 clients from the Bananas child-care agency, IFebruarv 1999 Exempt I Individual El Center |:| FCCH In Alameda, black and Latino clients chose exempt forms of child care more often than white clients. And they were less likely than whites to Choose center-based care. In addition, more than half of the Spanish—speaking clients chose exempt care as well as 43% of the English speakers. In Alameda County, fewer than half of the CalWORKs clients in our sample selected care in the same zip code in which they resided. Stage 1 clients. however, were more likely than Stage 2 clients to use providers in their home zip code. Los Angeles. In Los Angeles, both CalWORKs and APP clients opted for center—based care 39% of the time. CalWORKs clients, however, were more likely to choose exempt care while APP (non—CalWORKs) clients were more likely to use family child—care providers. We found strong associations between parent ethnicity and language and the type of care parents selected. White and Asian CalWORKs clients were more likely to pick center—based care (46%) than any other ethnic group. Latinos, on the other hand, selected exempt care more frequently than other groups (40%). Black CalWORKs clients in our sample were equally likely to chose center, exempt or FCCH-based care. Spanish—speaking CalWORKs clients chose exempt care more often than those who spoke English. Spanish—speaking and English- speaking APP (non—CalWORKs) clients, on the other hand, chose FCCHs approximately half the time. Age data for over 5,500 CalWORKs and APP children were obtained in September I999, and was again gath— ered for 3.500 children in December 1999. Using this data, we found that APP clients were more likely to choose FCCHs for infants across both panels. Although CalWORKs clients picked FCCHs for their infants 41% of the time in September, they were slightly more likely to choose exempt providers three months later. Among toddlers, we begin to see a rise in center-based care for both populations. The majority of CalWORKs populations selected center-based care for toddlers but APP clients continued to select FCCHS half the time. CalWORKs school—age children were more likely to be in exempt care while APP school—age children were more likely to be in center-based care. 18 Child care stability The final portion of this section examines change patterns for Alameda, Kern and Los Angeles County. By tracking particular indicators over time we were able to determine whether clients continued to receive child care subsidies and whether their providers and child-care type remained the same. This enabled us to infer the degree of stability experienced by subsidy recipients across the three counties. In Kern County. the tracking files were made up of the youngest child from each family using child-care subsi— dies. Of the CalWORKs and APP families using Stage 1 or 2 subsidies in March, 80% and 83%, respectively, appeared in the June data. So while there was some change in the type of care used by these families over the course of the tracking period, it was not a large difference. The percent of families using the same type of child care at all three points in time was also fairly high———86%. But while roughly 85% of the families were using the same provider in June and September, the retention rate between March and September had dropped to about 70% for both CalWORKs and APP clients, as Figure l2 shows. This suggests that providers were more likely to change over a longer time period. FIGURE 12 Child-care stability for CalWORKs and APP (non-CalWORKs) children in’Kern County: Share reporting same provider during March-September 199916 80% w 70%4 60% . 50% — l Center-based program 40% - I FCCHs 30% — El License—exempt 20% — 10% — .1A;‘c".‘;."£'a'w . . ‘ My... am“; .y CalWORKs children APP (non—CalWORKs) children % of children with same provider 0% In Alameda County, the tracking results showed less stability in child—care choice. Although 84% of CalWORKs children were in the same type of care from June to September 1999, the actual providers were the same for only 74% of the children. And by the next January, only 60% of the children were with the same child—care provider. Los Angeles’ indicators suggest that there was relatively high stability among both APP and CalWORKs clients. Matching across the months of September and December of 1999, 88% of the APP clients were still receiving subsidies. Of the APP clients, 98% were still using the same type of care and almost that percent still used the same provider. 19 CalWORKs clients in Los Angeles, however, showed less stability than their APP counterparts. Only 78% of the CalWORKs clients in September were still using subsidies in December. While the type of care chosen remained consistent at 95%, 83% of the Children were with the same provider. Despite this decrease in provider stability, the type of care chosen remained quite stable. More than 90% of children in September had the same type of care in December of 1999. Child Care Supply ISSUES In addition to understanding the characteristics and behaviors of the families served by the child-care subsidy system, it is important to ask whether licensed child-care organizations are expanding to keep pace with growing demand. Over the past five years, two California governors and the legislature have moved aggressively to expand the availability of licensed care, hoping to broaden parents’ access to formal centers, preschools, and FCCH homes which are regulated by the state. Under former governor Pete Wilson, funding increased for state preschools and child-care vouchers, primarily targeting welfare—poor and working-poor families. The Davis Administration, in the past three years, has focused on expanding the state preschool program and supporting center—based programs. Federal block—grant funding, since l990, has generally funded voucher—like mechanisms, providing support to both licensed organizations and indi- viduals who provide child—care services for low—income parents. Given this new funding, to what extent has the supply of licensed child care increased across the state? In general, we find that growth in child-care capacity—the number of children that a center or family child—care home is licensed to serve—is not keeping pace with the growing child population. Data from the California Child Care Resource and Referral Network in San Francisco shows that the capacity of center—based care for 2— to 5—year-olds rose slightly between 1996 and 1998 and then fell again in 2000. The picture for the capacity of family child—care homes is more promising. The capacity of these providers rose between 1996 and 1998 from 8.2% to 10.3% of the two child cohorts combined (ages 0-2 and 2—5). A portion of this apparent growth may have been a one—time gain, however, given that licensing rules were relaxed to allow for additional enrollment of school—age children in FCCHs. Capacity then leveled—out at the higher level between 1998 and 2000. 20 Figure 13 Change in California child-care capacity, 1996-2000 (medians across constant'sets of zip codes) % share of age cohort 25 — ,_ 20 _ . E11996 ——L ' E1998 ‘ 15 '. D2000 ~— 10 5 Center capacity, age 0—2 Center capacity, age 2-5 FCCH capacity Note: A constant set of zip codes is reported across the three time periods for each type of capacity: n=600 zips for center capacity for infants, age 0-2; n=919 zips for center capacity for preschoolers, age 2-5,- n=994 for family child-care home capacity, all ages. Zips with fewer than ’10 children age 5 and under or with no childcare capacity in 2000 were excluded. Data provided by the California Child Care Resource & Referral Network The capacity of centers that serve infants age 0—2 is quite low statewide. It did rise slightly, between 3.1% and 3.6% of this age cohort, in the 512 zip codes that have non-zero capacity and comparable data over the three time periods. But no gain in capacity, after factoring-in child population growth, is observable between 1998 and 2000. Note that these growth estimates are calculated using earlier child population estimates through the late 1990s, made by the Census Bureau. Initial analyses of year 2000 Census data shows that growth in California’s preschool—age population slowed by the mid-19905. Figure 14 displays the net change in capacity observed over the four—year period, 1996-2000. Center capacity for infants and toddlers climbed about one—half of one percent of this age cohort. Capacity for preschoolers actually declined slightly, failing to keep pace with child population growth. FCCH capacity grew between 1.5 and 2.0 percent of the age cohort. Note that mean differences show slightly larger changes, than medians, indicating that the distributions are somewhat skewed with certain zip codes growing or shrinking more markedly. 21 FIG‘UREV1'4i ‘Net change in California chilidi-carecavpaci’ty between 1996 and 2000(average changem capacity as share of age cohort) Change over 4 years in enrollment capacity as a % of age cohort [1 Net change medians a» Net change means '1 - “Center-capacity ageO-Z "Center'capacity age 2:5 ' '""F'CC'H“'capacity""“""“"“'" Note: A constant set of zip codes between 1996 and 2000: n=623 zips for center capacity for infants, age 0-2; n=955 zips for center capacity for preschoolers, age 2-5; n=1057 for family child—care home capacity, all ages. Zips with fewer than ’lO children age 5 and younger or with no child care capacity in 2000 were excluded. Data provided by the California Child Care Resource & Referral Network. Data from the licensing division of the California Department of Social Services provides a similar picture of statewide trends. Slots in licensed centers grew by 2.5% annually between 1997 and the first quarter of 2000.17 The number of new centers granted licenses grew at about two-thirds of this rate, suggesting that much of the expansion has occurred via newly created slots in existing centers. This holds equity implications, since we know that dispari— ties mark the distribution of center-based programs and preschools across different counties, social-class and ethnic groups statewide.Ix This 2.5% growth rate in center-based care is not impressive when placed in the context of child population growth (estimated at 3.4% per year during this period). The licensing data shows stronger growth in FCCH capacity, climbing at 3.3% annually after 1997, the three years following California’s welfare reform bill. But again, at least a part of this increase is due to the changes allowing more school—age children to be enrolled. The number of new applications for licenses was unchanged and the net number of operating FCCHs actually shrunk by 1.5% over the three—year period. These numbers include state and county—licensed jurisdictions. Comments from the CD88 licensing division clarify that new applications for FCCH candidates, excluding county—licensed jurisdictions, have crept upward slowly, rising from 809 monthly to 1,051 in 2000.“)This rate is slightly above child population growth, and it may be climbing more steeply since the first quarter of 2000, accord— ing to the licensing division. The growth or decline of child—care organizations is undoubtedly sensitive to local conditions. The number of center—based child slots in Los Angeles County, as reported by the Network, has grown at one—half of one percent in recent years. But annual growth in San Francisco’s aggregate center capacity has climbed at an annual rate of 5.7 22 percent in recent years.20 Beyond county differences, the Network’s data allow us to examine how capacity varies across differing types of zip-code communities, at any one point and over time. License-exempt care Any assessment of child-supply should also include a look at license-exempt group care programs, since licensed care represents only one option for parents who are seeking a supervised place for their children. License-exempt care includes kith and kin providers as well as group settings, such as the Boys & Girls Clubs, after-school pro— grams, parks and recreation programs, and sports leagues. A year ago, PACE began investigating how much license-exempt group care exists in Los Angeles County. Because there were no sources that collect this information, PACE conducted a survey of all the agencies and organizations that might offer license—exempt care. Over 400 surveys were sent out, 205 of which were completed. A precise accounting of the total amount of space available in license-exempt group programs is virtually impossible. Any count of slots or children served is, at best, an estimate because many of these programs accept children on a drop—in basis. many are part—day, after—school and weekend programs, and others come and go depend— ing on funding. Still, the survey provides at least a rough idea of how much of this form of care exists. Given those caveats, our estimate of the capacity of license-exempt group care programs countywide is 172 slots for infants, 4,800 slots for toddlers, and 55,000 slots for school—age children. Some of these operate part’iay, some are full—day, and some are summer or weekend only. Our estimate of the number of children, on average, participating in license—exempt group programs of any kind are 172 infants, 4,340 toddlers. and 39,900 school—age children. This estimate does not account for children who may be using more than one program, and it also reflects a very rough average of children served by those programs that provided a range rather than an absolute number. Comparing the overall capacity of license-exempt group care programs to licensed child—care in Los Angeles County, it is evident that these settings offer care to two very different populations. While licensed care in the county is weighted heavily toward preschoolers and toddlers—103,700 slots—there is relatively little toddler care in license—exempt group care settings—under 5,000. The opposite is true for programs for school-age children. License—exempt group care programs can potentially serve more than twice as many children as licensed slots in the county—55,000 vs. 25,000. Infant care is low in both categories, but again there is a dramatic difference. While 5,650 infants can be served by licensed child—care providers in the county, only 172 slots are available in license—exempt group care settings for infants. In conclusion, it appears that the greatest impact of license—exempt group care is in the area of school-age care. It potentially triples the available supply of programs for school—age children. Any plans for expanding care for school—age children should take this data into account. 23 Disparities in chlld—care SUDDIV and growth. Returning to the supply of licensed care, we’ve learned that access to licensed providers and centers depends on where one lives. While this report focuses on low-income families, it is also useful to place their communities in the context of better—off neighborhoods. Looking at center capacity levels across four income groups, we found that although capacity is at about 22% of the population of 2- to 5-year olds in low-income and working-class neighbor- hoods, the percentage increases to 24% in upper-middle class zip codes and to almost 30% in high-income areas. Therefore, even if there is growth in capacity, it is not clear where these additional slots will be created. Figure 15 i Child-care capacity by zip-code income quartiles, 2000 (mean capacity in each of four zip-code groups) Capacity as ”/0 of relevant age cohort ‘Il Center capacity, 2-5 ill? Family child-care home capacity O i—‘*“ Low—income zips Low—middle incomeHigh-middle income High-income zips Note: Zips sorted into four quartiles based on median household income in 1990. From the lowest to the highest quartile, average household income aggregated to zips equaled $22,010, $30,830, and $54,245, respectively. Zips with fewer than 10 children age 5 and under or with no child-care capacity in 2000 were excluded. Data provided by the California Child Care Resource & Referral Network. Capacity in FCCHs, however, does not show the same relationship with household income. Capacity is only slightly higher in zip codes populated by high—middle and high—income families, relative to zips within concentra— tions of lower-income households. These inequalities are starker when we contrast zip codes populated by parents who have a high school diploma with those who did not complete high school. The center capacity for preschoolers, ages 2 to 5, is more than twice as high for zip codes with adults who have a diploma (about 34% of the cohort in 1996), than it is for zip codes with the least educated population (15% of the cohort). We found some growth in capacity in zip codes with better—educated adults, but none in those zip codes where the level of educational attainment is much lower. 24 Figure 16 Child-care capacity, splitting zips by adult school attainment levels . i H . i Capacity as % share of age cohort E] Low schooling level (Q1) High schooling level (Q4) Center capacrty. Center capacity, FCCH capacity, FCCH capacity, 1996 2000 1996 2000 (Means across constant sets of zip codes) Note: A constant set of zip codes is reported for highest and lowest quartile of zip codes: n=465 zips for center capacity for preschoolers, age 2-5; n=502 for family child-care home capacity, all ages. Zips with fewer than 10 children age 5 and under or with no child-care capacity in 2000 were excluded. Data provided by the California Child Care Resource & Referral Network. Similar disparities are found when looking at FCCH capacity. Only the growth in capacity was greater in zip codes that had higher concentrations of less—educated adults. This latter finding is encouraging and may be linked to the capacity—building efforts of CD88 and CDE. We also found similar patterns when we sorted zip codes based on concentrations of Latino families. In predomi— nately Latino zip codes, the capacity of centers and FCCHs ranges below one—half the capacity observed in Zips with few Latino families. Growth in FCCH capacity was significantly greater, between 1996 and 2000, in Latino zips, compared to non—Latino zips. Here too, it would be helpful to know whether the CD88 and CDE’S capacity— building initiatives help to explain this growth. What community factors predict growth in licensed child-care capacity? It is useful to examine what economic and demographic factors encourage or inhibit growth in center—based programs and FCCHs. These fundamental forces may enhance or limit the state’s effort to boost the amount of licensed care. We examined whether these factors are associated with higher or lower enrollment capacity at a given point in time as well as over a span of time, 1996—2000. Using a multivariate regression model, we found that for 2— to 5— year-olds, center capacity was high in poor zip codes (as indicated by median household income), relative to blue- collar and lower middle—class zips. Center capacity rose, however, for upper middle—income and affluent zip codes. Statistical models appear in Appendix A. 25 With this pattern taken into account, the analysis showed that in zip codes with heavier concentrations of Latino families, center capacity was still low. In fact, for every 10% increase in the number of residents identified as Latino, there was a 3% lower enrollment capacity in that zip code. Finally, the share of mothers with preschool—age children who are employed is positively related to center capacity, after taking into account the other predictors. Turning to possible predictors of growth in center capacity, 1996-2000, we could pinpoint few factors that were significantly related. In general, this means that prior levels of center capacity are quite stable, relative to supply conditions in other zip codes around the state. Per capita capacity was lower in zip codes with higher rates of child population growth. Thus, any observed gains in center capacity often failed to keep pace with child population growth in high-growth areas. In addition, zip codes with stronger supplies of FCCHS displayed significantly higher growth in center—based programs. This may be tapping into an underlying variation in family demand for licensed care, especially zip codes with higher maternal employment rates. When we examined FCCH capacity using the same method, we found the opposite of what we observed with centers. FCCH capacity was low for poor zip codes, rose in middle-income areas, and then fell as the median household income increased. Very affluent zip codes had the lowest supply of FCCHs. This pattern may be the result of efforts to expand center-based supply in poor communities, which appear to discourage the growth of FCCHs. However, when communities are no longer eligible for targeted center support, FCCHs are necessary to respond to family demand for child care. And again, we found that zips with higher concen- trations of Latino families had lower FCCH capacity in 1996. We found that growth in FCCH capacity was lower in zips with higher maternal employment rates. This may indicate that demand for center—based care is higher in better-off zip codes. Or, in low—income zips, gains in mater- nal employment rates—resulting from strong labor demand and welfare reformw—have not resulted in larger FCCH capacity. In addition. similar to center growth, we found that FCCH capacity growth was lower in zips with stronger child—population growth. Again, FCCH providers simply can’t keep up in these rapidly growing communities. Continuing QUEStiOI‘IS The research findings in this paper provide information regarding the behavior of CalWORKs parents who access subsidies, as well as that of APP (non—CalWORKs) clients. While it begins to answer the questions of who is accessing subsidies, what kind of child care are they using, and how much licensed child care is available to parents across the state, it does not tell us why parents behave in the way they do, and why the supply of licensed child care is expanding as it is. These explanations are needed in order to develop optimal child—care subsidy and supply policies for low-income parents. Among the questions that need to be answered are the following: 26 at: Why are eligible CalWORKs participants not using child care subsidies? 2% What kind of care is being used by CalWORKs parents who are not using subsidies? as Why and how do parents choose their child-care providers? % Why is the supply of licensed child care growing in some areas, and not in others? Some of these questions will be answered, at least in part, by the work done by PACE researchers in parent focus groups. We have also completed a survey of current and former CalWORKs parents on their subsidy and child care choices and experiences. However. more work may be needed to develop a broader understanding of the child care options these parents have or perceive they have, as well as on the multiple factors influencing their behavior. In addition, it may be worth revisiting efforts to determine subsidy use rates with the new data that CDE is collecting on Stage 2 subsidies. More complete data on who is using both Stage 1 and Stage 2 subsidies, matched with better data on who is eligible (from the WTW25 forms) may allow for a better understanding of subsidy take up rates. Finally, with new supply growth data indicating that supply is increasing in rural counties while decreasing in urban counties. relative to the population of children age 0-5, it will be important to determine what factors are contributing to growth in some areas. and impeding it in others. Moreover, the reasons that more Family Child Care Homes have been opened in Latino versus non~Latino zip codes warrant investigation, as this is a positive trend. Past research has indicated that areas with large Latino populations typically have fewer child care centers than communities that are predominantly white. Whether the new homes resulted from state and local capacity—building initiatives or other events should be determined. 27 Appendix A. Table I. Predictors of child-care center'capacitv floriCaIi’f-‘ornia zip codes: 1996 baseline level and change over 1996-2000 Preschool-age capacity, children age 2-5 years old [weighted least-squares estimates, and t-statistics reported] Model 1 Model 2A Model 2B Predictors 1996 capacity Change in capacity over 1996-2000period per capita Without pop. growth With pop. growth Median household income -3.7 E-6 -6.l E—7 ~6.4 E7 (-2. l6)* (-0.73) (~0.77) Median household income 5.4 B] l 7.8 13-12 8.6 E-12 squared (2.87) ** (0.82) (0.91) % of mothers w/children 4.2 13—3 6.4 13-4 3.5 13-4 0—5 years, employed (3.43)*** (1.13) (0.64) % population. Latino ~3.1 E-3 -1.62 E-4 -9.8 13—05 (-10.41)*** (-1.15) (—0.73) Family child-care home -.03 .10 .10 capacity per capita, 1996 (-0.31) (2.31)* (2.46)* Center capacity per capita — —.07 -8.6 children 2—5 years. 1996 (-4.18)*** (-5.31)*** Child population growth. — _, -.70 annual rate. 1990—1999 (—4.84)*** Full equation Constant .30 .01 .02 (7.70)*** (0.56) (1.08) F-value 50.44*** 4.07*** 7.34*** N of zip codes 926 875 875 Adj. )3 .21 .02 .05 *p<.05, **p<.01, ***p<.001. Excluded from the regression models are percentage of families below the poverty line and adult education levels. both of which are highly collinear with median household income and maternal employment rates. 28 Table II. Predictors of family child-care home [FCCH] capacity for California zip codes: 1996 baseline level and change over 1996-2000 FCCH capacity, all child ages [weighted least-squares estimates, and t—statistics reported] Model 1 Model 2A Model 2B Predictors 1996 capacity per capita Change in capacity over 1996-2000 period Without pop. growth With pop. growth Median household income 3.6 B6 —1.2 E-7 —4.0 13—8 (6.28)*** (-0.33) (-0.11) Median household income 4.2 13—11 ~42 E—12 —5.8 E-l2 squared (—6.79) *** (-1.02) (-1.43) 9? of mothers w/children —3.7 13-4 -8.9 13-4 -l.0 13-3 .0-5 years. employed (0.84)*** (—3.18)** (-3.65)*** % population. Latino »l.5 E—3 -9.82 E-4 —7.4 E—05 (—l6.27)*** (—1.42) (—1.07) Center capacity per capita. ~35 E~3 2.82 15-3 -1.4 E—3 1996 (—0.33) (0.41)* (-0.20) Family child—care home — .22 0.2l capacity per capita. 1996 (l.ll)*** (l.00)*** Child population growth. — — -.29 annual rate. 1990—1999 (-4.02)*** Full equation Constant .06 .05 13—2 .05 (4.72)-t** (5.17) (5.23) F-value 95.96*** 7.02*** 8.42*** N of zip codes 949 934 934 Adj. 13 .33 .04 .05 *p<.05, **p<.01, ***p<.001. Excluded from the regression models are percentage of families below the poverty line and adult education levels, both of which are highly collinear with median household income and maternal employment rates. 29 Endnotes ' US. Census Bureau, January, 2002. http://www.census.gov/hhes/poverty/threshld/threshOO.html 3 Personal communication via e—mail, Sandi Nelson, Urban Institute, September 12, 2000, and Smith, K. (2000). “Who’s Minding the Kids? Child Care Arrangements.” Washington, DC: US. Bureau of the Census, p. 70 3 Acs, (3., & Loprest, P. with Roberts, T. (2001). Final synthesis of findings from ASPE’s leaver grants. Washington, DC: Urban Institute. J.— In Los Angeles, at the time of our data collection, Stage 1 child care was operated by the County Department of Public Social Services, while Stage 2 and 3 payments were managed by the county’s ten APP agencies. Stage 1 was in the process of being contracted out to the APP agencies. In Alameda County two of the APP agencies handled Stage 1 clients, while seven APP programs (including the two operating Stage 1 programs) served Stage 2 and 3 clients. In Kern County, one agency, operated by the County Office of Education, ran all stages of the Ca1WORKs and AP child-care subsidy programs. 'J‘ The temi “APP (non-CaIWORKs)” refers specifically to child care subsidies that are for low—income, working parents who are not CaIWORKs participants and that are funded separately from those subsidies targeted specifically for CalWORKs participants. Alternative Payment Program (APP) agencies do administer both Ca1WORKs and non-CaIWORKs child care subsidy programs, but both the funding sources and the regulations regarding who qualifies for subsidies differ between the programs. (7 Meyers, M.K., Heintze, T.. & Wolf. DA. (1999). “Child care subsidies and the employment of welfare recipients.” Berkeley: University of California, UC DATA. 7 Index constructed from CDSS county reports, CW—l 15 8 There are private child care subsidies, e.g., through private companies or foundations, which are not administered through APP agencies or other sources from which we could gather data. Thus, this comparison may somewhat underestimate the usage of subsidies. " Carroll, J. (2001). How to Pay for Child Care? Local Innovations Help Working Families. A PACE Policy Brief. Berkeley, CA: University of California, PACE. ”’PACE has recently completed a telephone survey on subsidy utilization issues in three counties, including Kern, in which current and former CaIWORKs participants were asked about their use. familiarity and experiences with the child care subsidy system. Initial reports from this study will be released shortly. “The initial sample for Stage 1 clients in Alameda County, selected in March 1999, was 449 clients. However, by the time we received our first data on these clients in June 1999, 143 had already been transferred to Stage 2, become inactive or been terminated. or simply disappeared from the files with no explanation. There was an equally sharp drop-off in numbers between June 1999 and September 1999. 13Constructed from CW—l 15 reports provided by CDSS Research & Evaluation Branch. I‘See also Jacobson. Linda. (2000) “Are Child Care Options Expanding?” A PACE Policy Brief. Berkeley, CA: University of California. Policy Analysis for California Education. MFuller, Coonerty, Choong, and Kipnis (1998). An Unfair Head Start: California Families Face UnequalAccess 20 Child Care. Berkeley, CA: University of California, PACE. ”Constructed from CW-115 reports provided by CDSS Research & Evaluation Branch ”Denominator is number ofclients using particular type of care in Panel 1 out of clients present in all three panels 17Thanks to Bill Jordan for assembling these data. Child Care Licensing Division (2000) “Day care work volume report (LIS 993—A).” Sacramento: California Department of Social Services. ”Fuller, Coonerty, Choong, and Kipnis (1998). An Unfair Head Starr. I" Bill Jordan, personal communication via e-mail, March 27, 2001. 20California Child Care Resource and Referral Network (2000) Child Care Portfolio. 2000. San Francisco: Network. 30 EEEEEEEEEEEEEEEEE WWW\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ CLI'HEH'MHH Polll' y AllleQIS for’ California Education PACE Unix/elem of Callfomua, Berkeley and Stanford Umiversitv 5653 Tolman Hall Belkelev, CA 947204670 Telephone. lSlOl 642—7225 Wm lullceberlelevedu