Comparing CalWORKs Take-Up Rates Across Demographic Groups

AN LAO REPORT Comparing CalWORKs Take-Up Rates Across Demographic Groups GABRIEL PETEK | LEGISLATIVE ANALYST | FEBRUARY 2023 Summary This is the second in a series of publications estimating the percentage of California Work Opportunity and Responsibility to Kids (CalWORKs)-eligible families who enroll in the program—otherwise known as the CalWORKs take-up rate. In the first post, we introduced our methodology for estimating the number of families eligible for CalWORKs since 2005 and compared this to the number who actually enrolled. We found that roughly 60 percent of eligible families enroll in CalWORKs. In this brief, we (1) compare the economic and demographic characteristics of our estimated CalWORKs-eligible population to those of the overall population of California, and (2) compare the CalWORKs take-up rates for various demographic and economic groups of eligible families. Generally, we find groups which historically have relatively high poverty rates are both more likely to be eligible for CalWORKs and more likely to enroll than those with relatively low poverty rates. These findings can inform future policy decisions aimed at increasing participation rates in the CalWORKs program. Comparing CalWORKs Eligibility are subject to error. That said, our estimates do Rates by Demographic Groups provide insight—particularly when comparing across groups—into CalWORKs participation. Two Surveys Provide Most of the Information We Will Continue to Review and Refine Our Used in This Post. Throughout the brief we rely Methods as More Data Become Available. on publicly available survey data to estimate the We are continuing to engage with knowledgeable characteristics of three populations: all Californians, stakeholders to refine the methodologies used CalWORKs-eligible Californians, and CalWORKs in this analysis. In particular, we are engaged in recipients. The textbox on the next page describes ongoing conversations with researchers within the the two surveys used throughout our analysis. Department of Social Services who are currently Because the COVID-19 pandemic has affected data conducting some similar analyses. Based on these collection for one of these surveys (and the state and conversations, and any results produced by the federal response to the pandemic has affected the administration, we may revisit some of the estimates availability of other economic supports generally and presented below at a later date. In addition, we CalWORKs participation in particular), the most recent intend to repeat the analyses presented below on year of data we use in this analysis is 2019. future data as it becomes available and report on All Estimates Provided in This Post Are Subject any major shifts we see. to Multiple Levels of Uncertainty. In particular, Some Demographic/Economic because we rely on survey data, our estimates are Characteristics Reflect Families Overall… subject to uncertainty regarding how representative The two surveys used for this analysis ask these surveys’ samples are of their underlying various questions about the attributes of the populations. As is always the case with survey data, respondents. In some cases, the questions identify this uncertainty is greater when we look at smaller and characteristics of the entire family. Examples of smaller components of the underlying populations. family-level characteristics include a family’s size Consequently, we advise caution in the interpretation (how many children it contains) or its composition of our estimates—in particular, the smaller the (whether it is headed by a single mother, single population estimated, the more likely our estimates father, two parents, or some other caregiver). www.lao.ca.gov 1 AN LAO REPORT Data Sources for Comparing Populations Data on California Work Opportunity and Responsibility to Kids (CalWORKs)-Eligible and Overall Populations Come From the American Community Survey (ACS). As described in our previous post, the ACS surveys a representative sample of more than 2 million households and contains sufficient economic data to allow us to estimate the number of individuals likely eligible for CalWORKs. In this analysis, we use these data to describe the overall California population as well as the population of CalWORKs-eligible families. The federal Census Bureau reports that COVID-19 significantly impacted data collection for the ACS in 2020. In addition, the pandemic saw the advent of many temporary supports for low-income families which may have affected CalWORKs enrollment in both 2020 and 2021. Consequently, the data reported from these years may not be comparable to that collected in prior years. Recognizing this, we will focus on data collected between 2005 and 2019 throughout this brief. We expect these data to best represent the CalWORKs-eligible and enrolled populations in a “typical” year. Data on the CalWORKs-Enrolled Population Come From an Annual Survey of Some Recipients. Federal law requires all states receiving Temporary Assistance for Needy Families funding (TANF, the federal funding source which covers some CalWORKs costs) to report certain demographic and economic data on recipients of that funding annually. In California, these data are collected through the Research and Development Exercise Project (RADEP), which surveys a random sample of approximately 3,000 CalWORKs recipients annually. Our Two Data Sources Are Not Perfectly Compatible. In some instances, the ACS and RADEP surveys ask slightly different questions about the same general information, requiring us to make some assumptions about how such data can be compared. In addition, the RADEP sample excludes CalWORKs families who do not meet certain eligibility requirements that would qualify the parents to receive TANF dollars. For example, parents which have exceeded the 60-month federal lifetime limit on TANF assistance are excluded from the RADEP sample (as are their children), even though state law allows children in such families to continue receiving CalWORKs assistance. Overall, the RADEP sample excludes about one-quarter of all CalWORKs cases. For purposes of this analysis, and due to data limitations, we have assumed these excluded families are sufficiently similar to other CalWORKs families to allow us to draw some general conclusions from the RADEP data. We have validated this assumption by referencing data from prior years, when all CalWORKs cases were included in the RADEP sample. However, if families currently excluded from the RADEP sample are no longer similar to other CalWORKs families, the results of this analysis would be different. Families also can be identified by the collective levels, and employment statuses. For example, a incomes of their members. The federal definition of family containing both Black and white members poverty is relative to family size, such that a single cannot accurately be assigned by the survey to a mother caring for one child while making $18,000 single racial characteristic. Such characteristics, a year would not (in 2019, the year we use for most thus, can only accurately describe individuals, of our analyses below) be considered impoverished, rather than families as a whole. Both surveys used whereas a single mother making the same for this analysis collect individual-level information amount but caring for two children would be. for all members of a household. …Whereas Others Reflect the Characteristics This Brief Explores Both Family and of Individuals. Other survey questions reflect only Individual-Level Characteristics. Below, we the characteristics of an individual. Many families use one approach when evaluating eligibility contain members of different races, educational and take-up rates by family-level characteristics, 2 LEGISLATIVE ANALYST’S OFFICE AN LAO REPORT such as family size, and a different approach between families headed by single mothers and when evaluating these rates by individual-level those headed by single fathers. characteristics, such as race. For family-level …And Also Based on Individual characteristics, we examine the CalWORKs Characteristics. Figure 2 on the next page shows eligibility and take-up rates of children, as all our estimates of poverty rates for parents in select CalWORKs families contain at least one eligible demographic groups in 2019. Our estimate for child but only about half of CalWORKs families the overall poverty rate for California parents is contain any eligible adults. (As noted in the box shown as a dotted line, whereas the poverty rates on page 2, parents of otherwise-eligible children for parents belonging to specific demographic may no longer be eligible for CalWORKs for groups are shown as blue bars. For example, reasons like reaching the 60-month time limit. The women had a higher poverty rate than parents as children of these adults, however, can continue a whole, whereas men had a lower poverty rate. to receive CalWORKs grants.) For individual-level In examining the historical data, we found groups characteristics, we examine the CalWORKs with above-average poverty rates in 2019 also had eligibility and take-up rates for adults, as adults above-average poverty rates in each year since are ultimately responsible for deciding whether to 2005 (when our analysis of CalWORKs take-up enroll their families in CalWORKs and thus their rates begins); the same is true for groups with characteristics seem most relevant in determining below-average poverty rates. take-up rates. Families Residing in Very Low-Income Poverty Rates Vary Based on Family-Level Households Are More Likely to Be Characteristics... To qualify for CalWORKs, CalWORKs-Eligible. CalWORKs eligibility is families generally must earn less than about determined based on family income, not household 80 percent of the federal poverty level. Figure 1 income. Thus, a multifamily household potentially shows our estimates of poverty rates for children by various family characteristics in 2019. Figure 1 Our estimate for the overall poverty Poverty Rates Vary Based on Family Characteristics rate for California children is shown LAO Estimates of California Children, 2019 as a dotted line, whereas the poverty rates for children belonging Poverty Rate for All Childrenª to specific demographic groups are Household Income Below $20,000 shown as blue bars. Bars extending Between $20,000-$75,000 beyond the dotted line represent Above $75,000 groups with relatively high poverty rates, whereas smaller bars Single Child represent groups with relatively Two Children Three Children low poverty rates. Unsurprisingly, Four Children because poverty is defined based Five or More Children on family size and income, poverty rates are strongly correlated with Two Parent both the number of children and Single Mother household income. In addition, Single Father Otherb single-parent households are far 20 40 60 80 100% more likely to experience poverty a LAO estimate from American Community Survey data. than two-parent households, b Other includes children living with other relatives such as grandparents or with non-relative caretakers. although there is little difference Note: All estimates subject to assumptions and methodologies laid out in original LAO report. www.lao.ca.gov 3 AN LAO REPORT could include both higher-income members and groups, with such households being particularly also a CalWORKs-eligible family. Figure 3 shows more common amongst Hispanic/Latino and the distribution of CalWORKs-eligible children by Asian Californians and particularly less common household income. Unsurprisingly, nearly every among white Californians. Upon closer inspection child residing in a very-low-income household of the available data, we believe this in part (those making $20,000 or less per year) is likely reflects a larger trend of CalWORKs-eligible CalWORKs-eligible, as are a smaller number of families moving in with relatively higher-income children residing in middle- and high-income friends and family. households. (Low-income families cohabiting with higher-income Figure 2 friends and family present a Poverty Rates Vary Based on Individual Characteristics particular challenge for our LAO Estimates of California Parents Residing With Own Children, 2019 estimate of the CalWORKs-eligible population, as discussed in the Poverty Rate for All Parentsª nearby box.) Male CalWORKs-Eligible Children Female Are Somewhat More Likely to Hispanic/Latinob Reside in Multigenerational White Households. Relative to all Black children in California, we find Asian CalWORKs-eligible children are High School or Less somewhat more likely to reside Some college in households including at least 5 10 15 20% one parent and one grandparent. Importantly, the frequency of such a LAO estimate from American Community Survey data. b We assign anyone with Latino/Hispanic ethnicity to this category regardless of race. This is consistent with how the multigenerational households administration reports CalWORKs data. differs between demographic All estimates subject to assumptions and methodologies laid out in original LAO report. Figure 3 CalWORKs-Eligible Children Typically Live in Very Low-Income Households LAO Estimates of California Households With Children, 2019 400,000 Not CalWORKs Eligible 350,000 CalWORKs Eligible 300,000 250,000 200,000 150,000 100,000 50,000 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100105110115120125130135140145150155160165170175180185190195200>$200 Dollars in Thousands Note: All estimates subject to assumptions and methodologies laid out in original LAO report. 4 LEGISLATIVE ANALYST’S OFFICE AN LAO REPORT Our Model Does Not Account for Contributions From Friends and Family. As discussed above, some low-income families live with higher-income friends or family members. An unemployed single mother, for example, might live with relatives. If this mother were to apply for California Work Opportunity and Responsibility to Kids (CalWORKs), current state regulations likely would require reporting the relatives’ gift of room and board as income. In some cases, such gifts might render this mother or similar applicants ineligible for CalWORKs benefits. (Although these in-kind contributions count as income for purposes of determining eligibility, they do not affect the size of grants for CalWORKs recipients.) Unfortunately, American Community Survey data do not allow us to easily simulate the effect of these regulations on potential CalWORKs applicants, and thus our eligibility model disregards all income from such gifts. Fully accounting for such gifts would cause some otherwise CalWORKs-eligible families to be considered ineligible, and this would cause our estimate of all take-up rates to increase. However, based on our understanding from county administrators and other stakeholders, we understand that applicants very rarely have such gifts applied against their income for eligibility purposes. Children From Groups Figure 4 With Relatively High Poverty Rates Are More Likely to Be CalWORKs Eligibility Rates CalWORKs-Eligible… Figure 4 Vary Based on Family Characteristics summarizes our remaining LAO Estimates for California Children, 2019 findings on CalWORKs eligibility by family-level characteristics. The Eligibility Rate for All Children overall eligibility rate for California Household Income Below $20,000 children is shown as a dotted Between $20,000-$75,000 Above $75,000 line, whereas the eligibility rates for children belonging to specific Single Child demographic groups are shown Two Children Three Children as blue bars. Generally speaking, Four Children we find CalWORKs eligibility rates Five or More Children are highest for children residing Two Parent in low-income households, and Single Mother increase alongside the number Single Father Othera of children in each household. 10 20 30 40 50 60 70 80 90% Combining this information with the a Other includes children living with other relatives such as grandparents or with non-relative caretakers. information from Figure 1—which Note: All estimates subject to assumptions and methodologies laid out in original LAO report. identified family characteristics correlated with above-average similar poverty rates. In reviewing the data, this rates of poverty—we find larger families and, appears to be because families headed by single not surprisingly, families with smaller incomes mothers are about twice as likely to report zero have both higher poverty rates and higher outside income as those headed by single fathers. CalWORKs-eligibility rates. Notably, however, (Families with two parents have lower rates of both families headed by a single mother are much more poverty and CalWORKs eligibility.) likely to be CalWORKs-eligible than those headed by a single father despite these families having www.lao.ca.gov 5 AN LAO REPORT ...As Are Parents From Groups With Relatively smaller families (or, at least, families containing High Poverty Rates. Figure 5 summarizes our relatively fewer CalWORKs-eligible parents and remaining findings on CalWORKs eligibility by children) are relatively more likely to enroll in individual-level characteristics. The overall eligibility CalWORKs than larger families. We discuss this rate for California parents is shown as a dotted line, finding in more detail in the nearby text box. whereas the eligibility rates for parents belonging Estimating Parental Eligibility Is Particularly to specific demographic groups are shown as Challenging... Estimating parental eligibility is blue bars. Again, combining this information with difficult because many parents are ineligible for the information from Figure 2—which identified CalWORKs due to factors not directly measured demographic groups with above-average rates of poverty—this figure shows that groups which have higher poverty Figure 5 rates also tend to have higher CalWORKs Eligibility Rates CalWORKs eligibility rates. For Vary Based on Individual Characteristics example, Black parents are more LAO Estimates for California Parents Residing With Their Children, 2019 likely to be CalWORKs eligible than white parents and also experienced a higher poverty rate. Eligibility Rate for All Parents Employed Comparing CalWORKs Unemployed Take-Up Rates by Not in Labor Force Demographic Groups Hispanic/Latinoa Take-Up Rates Differ Between White Families, Children, and Parents. Black Asian In our last post, we estimated the CalWORKs take-up rate among High School or Less families at about 60 percent. Our Some College estimates for the take-up rates of 5 10 15 20 25 30 35 40% CalWORKs-eligible children and parents are both lower than this a We assign anyone with Latino/Hispanic ethnicity to this category regardless of race. This is consistent with how the administration reports CalWORKs data. overall take-up rate. In both cases, Note: All estimates subject to assumptions and methodologies laid out in original LAO report. these lower take-up rates suggest Larger Families Are More Likely to Have Outside Income. As noted above, the fact that our estimates for parent and child take-up rates are below our estimate for family take-up rates suggests that larger families (or, at least, families with more California Work Opportunity and Responsibility to Kids [CalWORKs]-eligible parents and children) are less likely to enroll in the program than smaller families. One possible explanation for this is that larger families are more likely to have jobs or other sources of income, and (as we discuss below) families with outside income receive smaller CalWORKs grants. These smaller grants may dissuade some families with outside income from enrolling in the program. Among CalWORKs-eligible families, we find those with two eligible parents are more likely to report outside income than those with one or zero eligible parents. Similarly, CalWORKs-eligible families with three or more children are somewhat more likely to report outside income than those with only one or two. Accounting for these differences in outside income explains most of the gap between our estimated take-up rates for families, parents, and children. 6 LEGISLATIVE ANALYST’S OFFICE AN LAO REPORT by the American Community Survey (ACS). Such rate for all CalWORKs-eligible families (60 percent), factors include immigration status, compliance we estimate the take-up rate for families reporting with CalWORKs programmatic requirements, and no outside income is somewhat higher and lifetime limits on aid. Although we have attempted to those reporting at least some outside income is account for each of these factors, such adjustments much lower. are imperfect. To account for the effect of sanctions Overall, CalWORKs Take-Up Rates Are and lifetime limits, we used administrative data Highest for Groups With Relatively High Poverty on CalWORKs-enrolled families in which only Rates… Figure 6 summarizes our estimated the children are eligible for aid to estimate the take-up rates for parents across several economic corresponding number of ineligible parents. and demographic categories. The overall take-up (This reflects about half of CalWORKs families.) rate for CalWORKs-eligible parents is shown on This approach has a serious limitation: it cannot top. The dotted line provides a reference point account for families that chose to disenroll in to compare the take-up rate for all parents to CalWORKs after parents were rendered ineligible take-up rates for specific demographic groups. for aid. Accounting for such families would cause For example, we estimate women have a much our estimate of parent take-up rates to increase higher take-up rate than men. When comparing by an unknown amount. We therefore encourage Figure 6 to Figure 5 (showing relative eligibility extra caution in interpreting our estimates of rates) above, we find groups with relatively higher parent take-up rates, as these are likely to eligibility rates also tend to have higher take-up be underestimated. rates. For instance, Black parents are both more …But We Believe Comparisons Between likely to be CalWORKs-eligible than Asian parents Different Groups of Parents Are Still and also have higher take-up rates. Looking even Meaningful. For the purposes of this analysis, further back to Figure 2, we find that the groups we assumed there are no systematic differences with relatively high take-up rates also tend to in ineligible parents by demographic or economic be groups with relatively high poverty rates. For groups. To the extent this assumption holds example, Black parents have higher take-up and true, we can compare the relative eligibility and poverty rates than Asian parents. take-up rates of parents belonging to different demographic and Figure 6 economic groups. Thus, we can CalWORKs Take-Up Rates say that groups of parents with Vary Based on Individual Characteristics estimated take-up rates below the overall parent take-up rate LAO Estimates, 2019 are relatively less likely to enroll in the program, and those with Take-Up Rate for All Parents higher estimated take-up rates Employed Unemployed are relatively more likely to enroll in the program. However, to the Female extent that there are systematic Male differences between sanctioned Hispanic/Latinoª and timed-out adults and those White currently eligible for CalWORKs Black Asian across different demographic groups, these comparisons are High School or Less Some College less accurate. 10 20 30 40 50 60% CalWORKs Take-Up Rates Are Lower for Families With Outside a We assign anyone with Latino/Hispanic ethnicity to this category regardless of race. This is consistent with how the administration reports CalWORKs data. Income. Relative to the take-up Note: All estimates subject to assumptions and methodologies laid out in original LAO report. www.lao.ca.gov 7 AN LAO REPORT …With the Notable Exception of Hispanic/ why they do not enroll in CalWORKs is very difficult. Latino and White Families. In comparing Figures 2 In addition, our review of the relevant social science and 6, we generally find groups with above-average literature provided few insights. poverty rates also have above-average CalWORKs Despite these challenges, our conversations take-up rates. This is not the case, however, with stakeholders did yield some insights that for Hispanic and Latino parents, which have could be useful for beginning to understand why above-average rates of poverty but below-average families do not enroll. Moreover, our assessment rates of CalWORKs take-up. Nor is it the case for of program characteristics—in combination with white parents, which have below-average rates this stakeholder information—provide some of poverty but above-average rates of CalWORKs additional avenues for consideration. Although not take-up. Although there are likely many factors a comprehensive list, some possible reasons to which contribute to these results, one potential explore for take-up rate differences include: reason may be that Hispanic and Latino parents • Social Networks May Drive Program are the most likely to live in multi-generational Awareness. According to our stakeholder households and white parents the least likely interviews, most CalWORKs applicants first hear (according to our analysis of ACS data), and financial of the program from family members, neighbors, support from other family members could partially or colleagues. To the extent people live, substitute for CalWORKs assistance. work, or associate with others sharing similar Since 2005, CalWORKs Take-Up Rates economic and demographic characteristics, Have Decreased Across Almost Every Group. such word-of-mouth could result in “clusters” As discussed in our previous post, the overall of people enrolling in the program who share CalWORKs take-up rate for families has decreased these characteristics. from about 70 percent in 2005 to about 60 percent • For Some, Benefits May Not Be Worth the in 2019. This decline in participation occurred Effort to Get Them. In most situations, applying across nearly all demographic groups. Each of the for CalWORKs is a multistep process involving demographic groups examined above saw a relative relatively complicated paperwork and in-person decrease in their take-up rates between 2005 and interviews. Once enrolled, parents must continue 2019. The only exception we have found is among to submit regular reauthorization paperwork and, parents with more than a high school education, in many cases, must comply with ongoing work whose take-up rate increased modestly during or job-training requirements. Some parents may this period. feel these requirements are too burdensome Potential Reasons for Different and the associated benefits too low to justify enrolling in the program. This seems more likely Take-Up Rates Between Groups to be true of parents with outside income, for To Date, Relatively Little Work Has Been Done whom CalWORKs benefits are reduced relative on Understanding CalWORKs Take-Up Rates. to parents with no outside income. Specifically, a In preparing this brief, we talked to a broad selection family’s CalWORKs grant is reduced by 50 cents of stakeholders, including county administrative for every $1 above $500 per month they earn. staff, current and former CalWORKs recipients, This general cost-benefit analysis also may and groups providing guidance to CalWORKs explain why families with no eligible parents applicants. Although each of these stakeholders appear to have higher take-up rates, as families provided valuable insights into how CalWORKs with no eligible parents are exempt from meeting (and, especially, its application process) operates, work or job-training requirements. As prices for understandably, none claimed a deep familiarity housing, food, and consumer goods increase with the group of most interest to our research: at a more rapid pace than in recent years, this those families who are eligible for the program but cost benefit analysis may result in even fewer choose not to enroll. Unfortunately, because there eligible people deciding it is worth it to apply are few service ties to these families, understanding for benefits. 8 LEGISLATIVE ANALYST’S OFFICE AN LAO REPORT • Some May Be Relying on Support from Better Understanding Differential Take-Up Families Rather than CalWORKs Benefits. Rates Could Inform Future Policy Decisions. As we discuss above, some CalWORKs-eligible Policymakers interested in increasing CalWORKs families appear to live with relatively take-up rates may wish to target policy changes higher-income friends and family members. to groups with particularly low take-up rates. For at least some of these families, the financial For example, CalWORKs could be made more support they receive from social networks is attractive to families with outside income by likely serving as an alternative to CalWORKs. allowing these families to earn more without To the extent that some groups (for example, affecting their monthly CalWORKs benefits, or Hispanic and Latino or Asian Californians) by reducing the amount of paperwork these are more likely to live in multi-generational families have to complete to remain enrolled. households or otherwise cohabit with family or Increasing take-up for specific demographic groups friends, this could help explain why take-up rates might mean targeted promotional campaigns differ between groups. or re-branding efforts designed to reduce any • Negative Program Perceptions May potential stigma associated with the program. Discourage Enrollment. Another common We note that, as a result of a recent $2 million thread in our stakeholder interviews is that appropriation, the administration is currently in many CalWORKs applicants are reportedly the very early stages of developing a statewide very reluctant to apply for aid and do so only promotional campaign to increase program as a “last resort.” To the extent such attitudes utilization. In future posts, we hope to shed further vary between groups, this could explain some light on the potential efficacy of such strategies. differences in take-up rates between groups. www.lao.ca.gov 9 AN LAO REPORT 10 LEGISLATIVE ANALYST’S OFFICE AN LAO REPORT www.lao.ca.gov 11 AN LAO REPORT LAO PUBLICATIONS This report was prepared by Ryan Anderson, and reviewed by Ginni Bella Navarre and Carolyn Chu. The Legislative Analyst’s Office (LAO) is a nonpartisan office that provides fiscal and policy information and advice to the Legislature. To request publications call (916) 445-4656. This report and others, as well as an e-mail subscription service, are available on the LAO’s website at www.lao.ca.gov. The LAO is located at 925 L Street, Suite 1000, Sacramento, California 95814. 12 LEGISLATIVE ANALYST’S OFFICE