Report on the Outcomes and Characteristics of TANF Leavers

MARCH 15, 2017 Report on the Outcomes and Characteristics of TANF Leavers Carolyn Bourdeaux Lakshmi Pandey

Table of Contents Overview 2 Data and Methods in Brief 2 An Overview of Georgia s TANF Program, Policies, Economic Changes and Population Characteristics Historic Changes in the TANF Caseload 3 General Characteristics of TANF Recipients 7 Characteristics of the Adult Recipient TANF Leaver Population 11 Personal Responsibility Requirements 16 Short-Term Outcomes for TANF Leavers 16 Long-Term Outcomes for TANF Leavers 19 Escaping Poverty 23 Employment Outcomes for TANF Leavers by Industry and Wages 27 Conclusions 38 Endnotes 38 Appendix: Summary of TANF Expenditures 39 About the Authors 46 About the Center for State and Local Finance 46

2 Overview This report examines Georgia s Temporary Assistance for Needy Families (TANF) program, with a particular focus on the population that left TANF during and immediately after the Great Recession (2009-15). The report examines the characteristics of the TANF population as a whole and of TANF leavers as well as selected job market outcomes for those who left the program. This information updates an August 2007 report, Employment, Earnings and Recidivism among Georgia s TANF Leavers: Findings from the TANF Follow-Up System, produced by the Georgia Department of Human Services. Notably, the population described in this report is substantially smaller than that described in the earlier report. The population of interest, namely adults receiving benefits and subject to work requirements, only totaled around 2,274 as of May 2016 compared to 32,188 in January 2000. The earlier report focused on the relative effectiveness of state and federal mandatory work requirements. However, by the time of this report, these requirements have been largely assumed and institutionalized into the TANF program. For instance, the 2007 report identified a substantial population of TANF participants subject to mandatory work requirements but not participating; however, in the current program, no such population can even be identified except by the very few adults who are being sanctioned for noncompliance. As a result, this report is more abbreviated and largely focuses on characteristics of TANF leavers, the extent to which TANF leavers exit into the workforce, their workforce outcomes and the extent to which the TANF population is able to escape poverty in the three years immediately following exit. The report finds that the TANF program effectively moves people off and keeps people off of TANF. Compared to the early 2000s, spells on TANF are brief. As the economy has improved, a larger percentage of leavers exit into a job and fewer families are entering the TANF program. However, TANF leavers who remain in the Georgia labor force largely move into low skill jobs with wages well below the poverty threshold. TANF leavers seem to fare better if they are employed by a government entity or in the health services sector. Data and Methods in Brief The analysis draws on a detailed dataset provided by the Georgia Department of Human Services (DHS) describing TANF households and individual recipients. The report focuses on TANF leavers or closed cases. A case is considered closed if the household does not receive benefits for two successive months. Additionally, the analysis examines long-term outcomes by matching individual records in the TANF dataset with records from the Georgia Supplemental Nutrition Assistance Program (SNAP) dataset and records from the Georgia Department of Labor unemployment insurance (UI) system, which tracks employment and wages for individuals by industry type. Importantly, both of these datasets only capture Georgia outcomes, and the Department of Labor data only capture employment in businesses subject to UI requirements. Sectors of the economy that are excluded include federal employment and selfemployment. As noted in the 2007 report, this means that data on employment can be interpreted as the minimum level of employment and earnings among former TANF recipients (DHR 2007, 2). At the

199106 199203 199212 199309 199406 199503 199512 199609 199706 199803 199812 199909 200006 200103 200112 200209 200306 200403 200412 200509 200606 200703 200712 200809 200906 201003 201012 201109 201206 201303 201312 201409 201506 201603 Cases 3 time this analysis was developed, the Department of Labor data was only up to date through 2015; therefore, any longitudinal look at outcomes ends with 2014 to allow a full year of outcome data for 2014 leavers. (The data thus includes a full year of employment data for people who left TANF in December 2014.) Last, while this report largely follows the format of the 2007 report, we were not able to exactly duplicate the methodology for generating the caseload numbers in the earlier report. The numbers in this report are within several percentage points of the 2007 numbers but are not exact matches. An Overview of Georgia s TANF Program, Policies, Economic Changes and Population Characteristics HISTORIC CHANGES IN THE TANF CASELOAD TANF was created by the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996. It replaced the Aid for Families with Dependent Children (AFDC) program and shaped the new welfare system in the United States. The TANF program provides income assistance to needy families while also promoting job training, job preparation and work in order to move families to self-sufficiency. The program also focuses on preventing teen pregnancy and encouraging two-parent families (DHR 2012, iv). Since the law s enactment, public policies and economic conditions have significantly changed the number and types of families receiving welfare benefits in Georgia. Most notably, the caseload overall has dropped dramatically since the inception of welfare reform. In January 1997, Georgia had 116,254 households receiving TANF, compared to a mere 12,329 as of May 2016. Figure 1. Total Monthly Caseload, June 1991-May 2016 160,000 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 Year Month

200001 200007 200101 200107 200201 200207 200301 200307 200401 200407 200501 200507 200601 200607 200701 200707 200801 200807 200901 200907 201001 201007 201101 201107 201201 201207 201301 201307 201401 201407 201501 201507 201601 Cases 4 Georgia's TANF population includes two main types of cases: adult cases and child-only cases. Adult cases include an adult recipient who is the parent of a dependent child. With a few exceptions, such as when an adult cares for a child less than 3 months of age, adult recipients are subject to the TANF work requirements and lifetime limits on receipt of benefits. Adults in child-only cases, often a nonparent relative caring for a child, are not subject to the work requirements or time limits (DHR 2012, xiii); additionally, the poverty determination is associated with the child, not the family with whom the child is placed. Figure 2 shows monthly average caseloads for each type of case from January 2000 through May 2016. 1 Figure 2. Monthly TANF Caseloads by Type 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 Year and Month Total Child Only Cases Child and Adult Cases 1 Because of coding changes in the data, this report could not go back to the 1990s and pull comparable populations.

5 Table 1. Caseload Changes Jan 2000-May 2016 CHANGE START PEAK END START TO PEAK PEAK TO END START TO END Jan-00 Oct-03 May-16 Total Cases 55,984 59,937 12,329 7% -79% -78% Adult and Child 32,188 34,691 2,274 8% -93% -93% Child Only 23,796 25,246 10,055 6% -60% -58% Share Adult and Child 57% 58% 18% Share Child Only 43% 42% 82% During this period, the number of adult cases dropped from 32,188 to 2,274, a decline of 93 percent (see Table 1). Child-only cases dropped from 23,796 to 10,055, a decline of 58 percent. Obviously, the mix in the caseload also changed, with households with an adult recipient making up 57 percent of the total cases in 2000, but in 2016 only making up 18 percent of the caseload. To a large extent, TANF is no longer an income support program for traditional nuclear families but rather helps support nonparent or kinship care arrangements. The TANF program in Georgia officially began on Jan. 1, 1997. Georgia adopted a 48-month lifetime limit on the receipt of benefits (although hardship waivers were available in selected cases) and shifted case management emphasis to finding and retaining a job. This change in the program, combined with a strong economy, had the effect of halving the program from 116,213 cases in December 1996 to 50,996 cases by December 2000. Almost two-thirds (62 percent) of the adults leaving TANF appeared in the Department of Labor dataset reporting earnings in the first quarter after they exited the program (DHR 2007, 4). In January 2001, the first cohort of adult recipients reached their lifetime limit; however, only about 6 percent of the 2001-04 cohorts of leavers exited TANF because they had reached this limit. During this time, the country entered a recession, and the state faced declining employment in 2002 and 2003 (see Figure 3). The adult caseload began to tick back up, increasing to 59,605 by December 2003. During the 2001-04 period, only 55 percent of leavers recorded earnings in the UI system in the first quarter after they had left TANF (DHR 2007, 4). Figure 3 shows that in 2004 the economy began to add jobs. Also in 2004, the federal government notified the state that it would now be required to meet the 50 percent federal work participation requirement. 2 In response, Georgia initiated TANF = WORK NOW, a new service delivery strategy that emphasized educating TANF applicants about work requirements, required engagement in work activities within seven days of approval of the application, and closely monitored participants to ensure that they were adhering to the work requirements. This strategy, which was rolled out throughout 2004, resulted in 2 Prior to this, Georgia had not been required to meet this benchmark because of the significant and consistent declines in caseload. The policy was triggered by growth in the caseload (DHS 2007, 5).

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Non-Farm Employment 6 another significant drop in adult cases (DHR 2007, 5-6, 2012, iv). In January 2004 (see the vertical lines in Figure 2), the state had 33,906 adult TANF recipients, but by December 2006, this number had dropped by 87 percent to 4,459. The 2007 report on TANF leavers noted that at the beginning of 2004 approximately 28,000 recipients faced a mandatory work requirement but only 8,000 were actually participating. By the end of 2006, around 3,100 were subject to the mandatory work requirement and 2,200 were participating (DHR 2007, 6). Figure 3. Georgia's Annual Non-Farm Employment: 1990-2015 4,800,000 4,600,000 4,400,000 4,200,000 4,000,000 3,800,000 3,600,000 3,400,000 3,200,000 3,000,000 Year Although the 2007 TANF leaver report notes the countercyclical (or opposite of economic trend) nature of TANF, the program appears to have played only a minor role as a safety net during the Great Recession. Figure 3 shows the changes in Georgia s Annual Non-Farm Employment since 1990. During this time period, the country faced three recessions according to the National Bureau of Economic Research: from July 1990 to March 1991, from March 2001 to November 2001 and then from December 2007 to June 2009. Figure 1 shows that the TANF caseload rose visibly during the first two recessions. However, during the 2007-09 recession, Georgia saw no visible uptick in cases. Figure 4 is scaled differently to focus on the 2007-16 period. It shows an uptick in the adult population, which essentially doubled between June 2008 and December 2012; however, this increase was more than offset by the decline in child-only cases.

200712 200803 200806 200809 200812 200903 200906 200909 200912 201003 201006 201009 201012 201103 201106 201109 201112 201203 201206 201209 201212 201303 201306 201309 201312 201403 201406 201409 201412 201503 201506 201509 201512 201603 7 Further, the adult cases were sufficiently small to begin with that even a large increase still means caseloads are still negligible relative to the pre-2004 period (see Table 2). Figure 4. Impact of the Great Recession on TANF Caseloads Dec. 2007-May 2016 25,000 20,000 15,000 10,000 5,000 0 Total Child Only Cases Child and Adult Cases Table 2. Impact of 2001 Recession and 2007-09 Recession ADULT MIN JUN-08 ADULT PEAK DEC-12 CHANGE Total Cases 21,340 18,781-12% Adult and Child 2,514 5,228 108% Child Only 21,976 13,553-38% GENERAL CHARACTERISTICS OF TANF RECIPIENTS Table 3a breaks down the characteristics of TANF recipients households from 2009 to 2015, and Table 3b shows the same statistics for households in which an adult recipient is part of the case (the adult may be coded as sanctioned for some violation of program rules, such as non-compliance with work requirements, as well as a recipient). The numbers associated with households or cases in the top row of each table are average monthly counts. Total recipients and total people in households are totals for the year, two slightly different ways of looking at the population. Cases may come on and off of TANF over the course of a year; thus, even though the average caseload may be capturing duplicates, it accurately portrays the average workload for an agency. By counting recipients and total persons in cases, we avoid duplicates, but we also treat a person the same regardless of whether he or she is on TANF for a month or for the entire year.

8 Table 3a. Characteristics of Households on TANF from 2009 to 2015 CHARACTERISTICS 2009 2010 2011 2012 2013 2014 2015 CHANGE 2009-15 Households Receiving TANF (Avg. Monthly) 21,030 20,415 19,652 18,633 17,399 14,747 13,175-37% Households with Child-Only Recipients 85% 82% 80% 76% 76% 79% 80% -5% Total Recipients 37,921 37,519 36,716 35,920 33,937 28,774 26,087-31% % Children 91% 90% 89% 88% 88% 89% 90% -1% % Adults 9% 10% 11% 12% 12% 11% 10% 1% Average Benefit Amount $209 $210 $213 $215 $215 $220 $220 Total People in HH 93,480 90,107 86,729 82,398 77,510 66,838 61,395-34% % HH Residing in Metro 76% 77% 81% 83% 82% 82% 82% 6% % HH with Some Employment 47% 45% 45% 45% 47% 49% 51% 4% Average Number of Children 2.3 2.3 2.3 2.3 2.4 2.4 2.4 Median Age of Children 8 8 8 8 8 8 8 Head of Household Characteristics % Grandparent Head of Household 41% 40% 39% 38% 38% 40% 41% 0% % Female 95% 94% 94% 95% 95% 95% 94% -1% % Male 5% 6% 6% 5% 5% 5% 6% 1% % White 29% 29% 29% 29% 29% 29% 29% 0% % Black 70% 70% 70% 71% 70% 70% 71% 1% % Other Race 1% 1% 1% 1% 1% 1% 1% 0% % Married 27% 27% 26% 26% 26% 26% 26% -1% % Never Married 52% 53% 54% 56% 55% 54% 54% 2% % Other Marriage Status 21% 20% 20% 19% 19% 20% 21% 0% Median Age 45 45 44 44 44 45 46 % With High School Degree/Diploma 46% 47% 47% 49% 50% 52% 53% 7% Households with Family Cap 848 821 808 776 693 569 465

9 Table 3b. Characteristics of Households with Adult Recipients (or Adults with Sanction), 2009-15 CHARACTERISTICS 2009 2010 2011 2012 2013 2014 2015 CHANGE 2009-15 Households (HH) Receiving TANF (Avg. Monthly Caseload) 3,526 3,980 4,295 4,702 4,418 3,338 2,824-20% % HH with Sanction 9.36 8.45 7.70 6.34 5.79 6.52 6.76 % Receiving SSI 4.78 4.78 4.48 4.71 4.77 5.57 6.74 Recipients 9,745 10,974 11,942 13,400 12,814 9,887 8,579-12% % Children 66.89 66.54 66.51 66.85 67.28 68.16 69.10 % Adults 33.11 33.46 33.49 33.15 32.72 31.84 30.90 Average Benefit Amount $284 $279 $281 $278 $278 $288 $285 Total People in HH 14,682 15,700 16,868 18,362 17,483 13,474 11,747-20% % HH Residing in Metro 85.63 85.75 88.06 89.80 89.03 87.34 90.68 % HH with Some Employment 32.19 30.16 31.97 32.38 35.90 38.46 41.92 Average Number of Children 2.5 2.5 2.5 2.6 2.6 2.7 2.8 Median Age of Children 3 3 4 4 4 4 4 Heads of Household (HoH) Characteristics (percent) % Female 97.93 97.88 97.75 97.80 98.09 97.94 98.06 % Male 2.07 2.12 2.25 2.20 1.91 2.06 1.94 % White 17.40 15.44 14.62 13.63 14.13 14.13 13.62 % Black 81.44 83.52 84.48 85.41 84.79 84.69 84.97 % Other Race 1.16 1.04 0.90 0.96 1.08 1.18 1.42 % Married 10.81 10.00 8.96 8.33 9.23 9.46 8.60 % Never Married 81.22 82.97 84.02 85.09 84.51 83.93 84.44 % Other Marriage Status 7.97 7.03 7.02 6.57 6.26 6.61 6.96 Median Age 25 25 26 26 26 27 28 % With High School Degree/Diploma 55.73 55.92 56.39 57.85 59.62 61.79 62.97 # Family Cap 28 29 34 35 27 22 21

10 In 2009, the total TANF average monthly caseload was 21,030. By 2015, it had declined by 37 percent. The average adult recipient caseload was 3,526 in 2009, dropping by 20 percent to 2,824 in 2015. Additionally, the average monthly caseload continued to decline over the 2009-12 period, whereas the adult caseload rose by 33 percent over the same period. However, by 2015, the adult caseload had declined by 40 percent to 2,824 cases, close to Georgia s historic low of 2,514 in June 2008. The demographics of the total population of recipients looks different from the adult case population because of the different considerations associated with child-only placements (i.e., no work requirement, no lifetime limit, and when a kinship care arrangement is involved, income is tied to the child s family not the host family). In the total population, the median age of children is 8 compared to 4 for the adult recipient cases. In general, a member of the household is more likely to be employed in the total population than in the adult recipient cases (e.g., 51 percent vs. 42 percent in 2015), and the head of the household is more likely to be married (e.g., 26 percent vs. 9 percent in 2015) and to be substantially older (e.g., 46 vs. 29 in 2015). Interestingly, the general TANF population head of household is less likely to have a high school degree than the cases in which the adult receives some benefits (53 percent vs. 63 percent in 2015). While there has long been concern about education levels of those in poverty, a high school diploma may no longer be enough to pull most families out of poverty. In 2009, 56 percent of adult TANF recipients had a high school diploma, rising to 63 percent in 2015 and this is over a smaller population of recipients with all the benefits of a more robust economy. We further examine these numbers in later sections of this report, but one area for inquiry beyond the scope of this report may be the type of education required to ensure that families can both find jobs and move out of poverty. Some additional geographic and demographic characteristics of the TANF population are interesting relative to overall numbers in Georgia. First, in 2009, 86 percent of adult heads of household receiving some form of TANF benefit were located in the 29-county Metro-Atlanta area, increasing to 91 percent by 2015. 3 For the total TANF population, 82 percent resided in the Metro-Atlanta area in 2015. However, in 2015 the 29 Metro-Atlanta counties only comprised around 55 percent of the state s overall population (5.7 million out of 10.3 million). Similarly, there are some interesting demographic differences in poverty and access to TANF. The white population made up 26 percent of the population in poverty in 2015 in Georgia but represented only 14 percent of the adult cases receiving some form of TANF and 29 percent of total households receiving benefits. The black population in Georgia made up 53 percent of the population in poverty, but black heads of household represented 85 percent of the cases in which adults received benefits and 71 percent 3 Metro counties are the 29-county Atlanta-Athens-Clarke County-Sandy Springs Combined Statistical Area (CSA). Note that Hall County is dropped from the TANF caseload calculation because it was added by the US Census in 2010. The total population in the metro area is 5.7 million (including Hall County) out of a population of 10.3 million in Georgia per US Census figures (metro area released March 2016). Thus, the 2009 and 2015 numbers are not perfectly comparable; however, the difference is small. The 2016 population estimates are as of July 1, 2016.

11 of the total households receiving benefits. This difference in percentage share of poverty and percentage enrollment in TANF appears to stem in part from very low representation from any other group. In 2015, Hispanic and other designations make up 20 percent of the households in poverty in Georgia, but only 1 percent of other race cases receiving TANF benefits (Kaiser Family Foundation 2015). These demographic variations should be considered in the context of a program that is extremely small at this point. TANF only addresses the needs of a very limited proportion of the population in poverty. According to the US Census Bureau, 1.8 million Georgians were living in poverty in 2015. At most, based on TANF data, a total of 26,087 people were receiving benefits only 1 percent of the population in poverty and of these, 90 percent were children. Keeping in mind the very modest population being served, the mismatch between recipients, population centers and populations in poverty suggests that further investigation into issues of access to and knowledge about the TANF program may be needed. CHARACTERISTICS OF THE ADULT RECIPIENT TANF LEAVER POPULATION This section focuses on households with adult TANF recipients that left TANF between 2009 and 2015. 4 Adults are subject to mandatory work requirements, time limits, and other restrictions and requirements around the receipt of benefits. 4 Adult cases are coded as such when an adult is coded as a recipient or is coded as sanctioned for non-compliance. When determining whether a household is a leaver, we assumed that a one-month break is a data issue and filled in the missing data with the previous month s value. If the gap is two months or more, we considered the household to have exited the program. Final exit means that the case did not receive any benefits after that month.

12 Table 4. TANF Leavers Characteristics for Households with Adult Recipients DESCRIPTION 2009 2010 2011 2012 2013 2014 2015 Total Households (HH) Exiting 4,520 5,036 5,450 5,897 6,211 4,995 4,669 % HH with Sanction 4.91 4.59 4.42 3.29 2.82 2.88 2.51 Total People in HH 16,776 18,281 19,634 21,470 22,829 18,621 17,673 % of People in HH Receiving Benefits 76.39 77.20 78.74 79.31 79.85 80.29 81.55 Average Number of Dependent Children 1.8 1.8 1.8 1.9 1.9 2 2.1 Median Age of Youngest Child 1 1 1 2 1 1 1 Heads of Household (HoH) % Female HoH 97.69 97.93 97.76 98.01 98.19 98.49 98.52 Median Age of HoH 26 26 26 26 26 27 27 Race of HoH % White 22.67 19.63 18 16.43 15.92 14.51 12.65 % Black 76.3 79.07 81.04 82.34 82.94 84.22 85.94 % Others 1.01 1.29 0.95 1.22 1.12 1.26 1.39 Marital Status of HoH % Married living together or apart 12.69 11.1 10.51 9 8.74 8.86 7.86 % Never married 79.15 81.41 82.78 84.7 85.17 84.8 85.71 % Divorced, Separated, Widowed 8.14 7.48 6.69 6.29 6.08 6.32 6.42 % With High School Degree/Diploma 60.92 62.29 62.31 62.26 63.85 64.6 66.88 Median Accrued Lifetime Months of Benefits 6 7 7 7 8 9 8 % Received more than 48 months of benefits 2.5 2.4 2.2 2.22 1.99 2.22 2.07 % Resided in Metro Counties 81.88 83.24 85.7 89.09 89.71 87.76 91 Employment in Exit Quarter % HH with some employment 46.81 47.83 47.33 49.73 53.4 56.77 60.74 % HoH with employment 40.19 41.67 41.81 44.7 48.14 52.01 55.72

200901 200904 200907 200910 201001 201004 201007 201010 201101 201104 201107 201110 201201 201204 201207 201210 201301 201304 201307 201310 201401 201404 201407 201410 201501 201504 201507 201510 13 Table 4 shows that in 2015, 4,669 households with an adult welfare recipient left TANF. While this reflects a 25 percent decline from 2013, the number is very close to the number of cases leaving the TANF program in 2009. Table 3b shows that over this same 2009-15 period, the average monthly caseload for adult recipients declined by 20 percent. This finding suggests that a larger proportion of the adult caseload was leaving TANF in 2015 than did in 2009, an observation further supported by continuing declines in the average monthly caseload in 2016. The May 2016 monthly average caseload was 2,274 adult recipient cases. Figure 5 shows the percentage of heads of households who entered and exited the TANF program with a job; during the latter half of 2015, more than 60 percent of those exiting the program had a job. Note that final exit means that the case did not receive any benefits after the month of exit. Figure 5. Entry to and Final Exit from TANF for Cases with an Adult Recipient (2009-15) 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 % HoHH Starting TANF with jobs % HoHH Exiting TANF with jobs A comparison of Tables 3b and 4 clearly shows that, on average, adult recipient cases exiting the program had fewer children than the average adult recipient case (1.8-2.1 vs. 2.5-2.8 for the median average of all adult cases) and the median age of the child is younger (1 vs. 3-4 in all adult cases). Because adults are exempt from work requirements during the first three months of an infant s life, it is possible that a number of adults receive TANF benefits during this period and then quickly rotate off of TANF when the work requirement kicks in. Interestingly, the proportion of heads of household leaving TANF with jobs grew from 40 percent to 55 percent between 2009 and 2015, a 15 percentage point increase. During this same time period, most households with adult leavers were also more likely to have someone in the household who was employed, jumping from 47 percent in 2009 to 61 percent in 2015. The number of leavers with a high

14 school degree/diploma increased from 61 percent to 67 percent, and the percentage of those leaving with a high school degree/diploma is consistently higher than the same statistic for the average adult caseload in general (Table 3b), which ranges from 56 percent in 2009 to 63 percent in 2015. Table 5. TANF Leavers Characteristics for Households with Adult Recipients 2005 versus 2015* DESCRIPTION 2005 2015 Total Households (HH) Exiting 26,996 4,669 Average Number of Dependent Children 1.9 2.1 Median Age of Youngest Child 2 1 Heads of Household (HoH) % Female HoH 96 98.52 Median Age of HoH 26 27 Race of HoH % White 19 12.65 % Black 81 85.94 % Others <1 1.39 Marital Status of HoH % Married living together or apart 6 7.86 % Never married 76 85.71 % Divorced, Separated, Widowed 18 6.42 % With High School Degree/Diploma 62 66.88 Median Accrued Lifetime Months of Benefits 14 8 % Resided in Metro Counties 79 91 % HoH with employment 54 55.72 *Numbers may be slightly non-comparable because of differences in data coding between 2005 and 2015. Table 5 compares TANF leaver characteristics in 2015 and 2005, using data from the 2007 report (DHR 2007, 9, Table 1). While the numbers may not be a precise match due to differences in aggregating and coding the data, they are generally close. Some of the statistics are noticeably different. For example, the total number of leavers was dramatically higher in 2005 at 26,996 compared to just 4,669 in 2015. The median accrued lifetime months of benefits dropped from 14 months in 2005 to eight months in 2015. Leavers (and participants generally) are increasingly likely to live in the Metro-Atlanta area than elsewhere in the state, with 91 percent of leavers in metro Atlanta in 2015 compared to only 79 percent in 2005. Finally, the proportion of leavers who were never married rose from 76 percent in 2005 to 86 percent in 2015. Despite these differences, the percentage leaving with a job has stayed relatively stable at 54-55 percent, and the percentage with a high school degree/diploma was only slightly lower in 2005 (62 percent) than in 2015 (67 percent).

15 Table 6. 2012 Adult Recipient Leavers: Characteristics by Accrued Lifetime Benefit Months MONTHS 1-12 13-24 25-36 37-47 48+ Total Households (HH) 4086 1060 400 240 111 % of Total Households 69% 18% 7% 4% 2% Total People in HH 13,481 4,344 1,869 1,161 615 % of People Receiving Benefits 83.48 76.31 73.24 68.04 48.61 % Female HoH 97.79 98.3 98.75 100 96.39 Median Age of HoH 25 26 29 31 35 Race of HoH % White 18.84 12.07 10.75 7.08 9.9 % Black 79.78 86.6 89.25 92.08 90.09 % Others 1.37 1.32 0.83 Marital Status of HoH % Married living together or apart 9.05 7.07 10.75 10 17.11 % Never married 84.94 86.41 81.75 85 69.36 % Divorced, Separated, Widowed 5.99 6.5 7.5 5 13.51 Average Number of Dependent Children 1.7 2.1 2.4 2.3 2 Median Age of Youngest Child 1 2 3 4 5 % HoH with High School Degree/Diploma 64.14 59.62 55.5 58.33 51.35 Median Accrued Lifetime Months of Benefits 5 17 29 43 51 % Resided in Metro Counties 88.81 90.75 87.25 90 88.28 Employment in Exit Quarter % HH with some employment 49.87 49.81 51.25 44.58 49.54 % HoH with employment 46.37 44.05 43 32.5 21.62 Table 6 looks at 2012 leavers by their accrued lifetime months of benefits. This snapshot, combined with the Table 5 findings, suggests that the nature of TANF has changed from the period examined in the 2007 report. Table 6 indicates that the heads of household on TANF are mothers with young children. It appears that they might be using TANF for support during the early days of the child s life and then leaving TANF relatively quickly as the work requirements kick in: 69 percent of leavers had a child with a median age of 1 and had only five median accrued lifetime months of benefit. In contrast, in 2005, only 45 percent of leavers had less than 12 months of accrued benefits (DHR 2007, 12). At the other end of the distribution, only 13 percent of TANF leavers had been on TANF for more than two years (in 2005, it was 31 percent). These households had older parents and older children and the heads of household were less likely to have a high school degree or diploma. However, the work outcomes are only marginally different: from one to 36 months, about 50 percent of leavers exited to employment. Only 45 percent of leavers who had been on the program for 36 months or longer exited to employment. A very select group appear to have received benefits beyond the 48-month state lifetime

16 limit. This group would have to have some extenuating circumstance such as domestic violence that allows the state to extend the benefits in the first place. Interestingly, exit into employment for this population was also about 50 percent. PERSONAL RESPONSIBILITY REQUIREMENTS One condition of receiving TANF is that the state may require a recipient to participate in various life skills activities. Table 7 shows the range of activities and the average months required to participate as a percentage of a participant s total months on TANF. Georgia appears to focus on topics around parenting, including participating in school-based parent-teacher conferences, parenting skills classes and family planning counseling. In 2014 and 2015, the state appears to be have begun focusing more substantially on life skills classes. In 2013, adults participated in life skills classes for 35 percent of months they were on TANF; by 2015, this number had jumped to 50 percent. Table 7. Personal Responsibility Requirements (Adult Cases) AVERAGE NUMBER OF MONTHS REQUIRED AS A PERCENTAGE OF TOTAL NUMBER OF MONTHS ON TANF 2009 2010 2011 2012 2013 2014 2015 Parent Teacher Conference 37.79 39.00 37.27 36.27 38.77 41.08 41.70 School Attendance 4.38 3.93 3.10 2.95 2.95 2.77 3.36 Parent Skill Classes 70.79 72.61 71.06 70.47 73.19 77.13 77.46 Family Planning Counseling 38.55 39.91 38.33 36.87 39.49 41.67 42.16 Financial Management Counseling 4.73 4.62 4.31 3.93 3.44 2.37 2.32 Life Skills Class 36.61 36.12 36.50 34.33 35.30 41.15 50.32 Substance Abuse Counseling 5.76 6.70 4.56 3.54 2.76 1.69 1.48 Mental Health Counseling 3.40 2.79 2.18 2.03 1.70 1.12 1.02 Rehabilitation Services 2.24 2.15 2.02 1.51 1.14 1.33 1.42 DFCS Child Welfare Case Plan 1.98 1.52 1.44 1.26 1.14 1.10 1.01 Work Plan Required 1.37 1.15 1.08 1.24 0.85 0.87 1.13 5.59 4.82 4.54 4.01 3.14 3.29 3.56 Short-Term Outcomes for TANF Leavers We next focus on adult leavers employment and recidivism in the first year after exiting TANF. According to the 2007 report, The TANF work requirements, combined with lifetime limits on receipt of benefits, reflect a belief that most recipients can make the transition to steady employment capable of supporting themselves and their families, within the benefit period. The employment, earnings, and recidivism of TANF leavers over the first year after they leave TANF reveal the efforts of former recipients to meet the challenge set by TANF policy, and the extent to which they are successful (DHR 2007, 17).

Table 8. Employment and Return to TANF in the First Year of Exit (Adult Recipient Cases by Annual Leaver Cohort) 2009 2010 2011 2012 2013 2014 17 % COHORT QTLY. WAGE % COHORT QTLY. WAGE % COHORT QTLY. WAGE % COHORT QTLY. WAGE % COHORT QTLY. WAGE % COHORT QTLY. WAGE Total Cohort Cases Exiting 6,006 6,714 7,207 7,576 7,622 5,785 % Employed, Did Not Return Within a Year 59.3 $3,027 58.9 $3,023 58.1 $3,033 59.5 $3,074 63.8 $3,102 68.9 $3,188 % Employed, Returned Within a Year 5.5 $1,913 6.4 $2,020 7.7 $2,126 9.1 $2,010 8.5 $2,116 7.9 $2,075 % Not Employed, Returned Within a Year 4.5 4.5 5.5 4.5 3.8 2.6 % Neither Employed Nor Returned Within a Year 30.8 30.2 28.8 26.9 23.9 20.6 % Employed in the First Year after Exit Quarter 64.7 $2,933 65.3 $2,925 65.8 $2,927 68.6 $2,933 72.3 $2,987 76.8 $3,074 % Those Employed in Exit Quarter 44.0 $3,511 43.4 $3,517 43.3 $3,536 46.3 $3,481 50.1 $3,547 54.0 $3,592 % Those Not Employed in Exit Quarter 20.7 $1,708 21.8 $1,749 22.5 $1,756 22.2 $1,790 22.2 $1,723 22.8 $1,846 % Total Returned 9.9 10.9 13.2 13.6 12.2 10.5 % Those Employed in Exit Quarter 3.4 $2,204 3.8 $2,047 4.8 $2,451 5.8 $2,370 5.4 $2,408 5.5 $2,330 % Those Not Employed in Exit Quarter 6.5 7.0 8.3 7.8 6.8 5.1

18 Table 8 shows the recidivism rates within the first year after exiting for leavers that were and were not employed when they left the TANF program. Fewer people returned to TANF in the 2009-14 period than in the 2002-05 period covered in the first report. In general, only 10 percent (in 2009) to 13.6 (in 2012) percent of leavers returned within a year, despite this period covering the depths of the Great Recession. In contrast, the population returning ranged from 27 percent in 2002 to 15 percent in 2005 (DHR 2007, 17). The recession s effects are more apparent in the percentage of leavers without a recorded job than in recidivism rates. In 2009 and 2010, around 30 percent left without a job and did not return. By 2014, this number had dropped to around 20.6 percent. Generally, job prospects improved between 2009 and 2014, although quarterly wages remained essentially the same. In 2009, 59.3 percent of leavers had a job and did not return, whereas in 2014, 68.9 percent had a job and did not return despite the overall number of cases also declining, presumably leaving a population that is harder to place. To calculate the wages shown in Table 8, we examined the four quarters after exit for any earnings, added these up and divided by four; thus, the calculation likely captures periods of partial employment. During the 2009-14 period, the quarterly wages of TANF leavers who had a job and did not return grew slightly from $3,027 to $3,188, in nominal dollars. Approximately 5-7 percent of the total cohort cases were employed when they left but returned to TANF within a year. Not surprisingly, quarterly earnings are lower for this segment of the population, but it is impossible to determine whether they returned because the salary was low or because they dropped out of the labor market and this affected their quarterly returns. Generally, being employed right away when leaving TANF appears to be associated with a more favorable long-term outcome. In 2009, of the 64.7 percent employed a year after exit, 44 percent were employed in the exit quarter. In 2014, of the 76.8 percent employed a year after exit, 54 percent were employed in the exit quarter. Of those who returned within a year in 2014, about half were employed in the exit quarter, but the numbers are small with only 10 percent returning. In sum, a solid majority of leavers seem to be employed, and many seem to retain their jobs at least for the first year. This job situation has improved in step with the economy, with more families leaving for employment rather than just leaving. However, Table 8 also shows that during the economic downturn, families exited to no clear employment, and most who exited, regardless of whether they were employed, did not return. For all of these families, quarterly earnings are generally quite small, with average quarterly earnings a little above $3,000. This translates into an annual income of around $12,000. For a family of two, the poverty income threshold is $16,020. For a family of three (a parent and two children), the poverty level is $20,160. 5 Median family income in Georgia is $47,829 according to the 2013 US Census American Community Survey. 5 https://dch.georgia.gov/sites/dch.georgia.gov/files/2016_federal_poverty_guidelines.pdf

19 Long-Term Outcomes for TANF Leavers Table 9. Employment, Earnings and Recidivism by Annual Leaver Cohort and Post-Exit Year* 2009 2010 2011 2012 2013 2014 FIRST YEAR AFTER EXIT Total Exiting 6,006 6,714 7,207 7,576 7,622 5,785 Employed in at Least One Quarter 64.61 65.14 65.58 68.5 72.23 76.73 Employed in All Four Quarters 30.5 30.95 30.45 33.8 37.58 39.6 Ave. Annual Earnings (Employed Only) $10,191 $10,154 $10,070 $10,273 $10,580 $10,890 Ave. Earnings per Quarter Worked $2,849 $2,825 $2,816 $2,871 $2,917 $2,973 Returned to TANF in First Year 9.92 10.85 13.15 13.62 12.24 10.54 Ave. Months on TANF (Returned Only) 5.74 6.69 6.75 6.89 7.77 8.03 Neither Employed Nor Returned 30.81 30.23 28.77 26.88 23.91 20.57 SECOND YEAR AFTER EXIT Employed in at Least One Quarter 62.72 62.25 65.28 70.39 73.58 Employed in All Four Quarters 32.21 31.66 33.63 38.42 40.93 Ave. Annual Earnings (Employed Only) $10,982 $10,984 $10,930 $11,583 $11,890 Ave. Earnings per Quarter Worked $3,114 $3,125 $3,109 $3,233 $3,310 Returned to TANF in Second Year 6.47 7.19 6.2 5.2 5.01 Ave. Months on TANF (Returned Only) 5.43 7.05 6.75 6.80 8.03 Neither Employed Nor Returned 35.04 34.85 32.71 28.07 25.21 THIRD YEAR AFTER EXIT Employed in at Least One Quarter 62 64 68.05 72.95 Employed in All Four Quarters 33.03 34.36 38.75 42.77 Ave. Annual Earnings (Employed Only) $11,928 $11,780 $12,098 $12,969 Ave. Earnings per Quarter Worked $3,344 $3,296 $3,348 $3,564 Returned to TANF in third year 3.69 3.32 2.63 2.38 Ave. Months on TANF (Returned Only) 5.54 6.22 6.87 6.56 Neither Employed nor Returned 36.51 34.85 31.24 26.54 *Adult recipient cases (or adult has been sanctioned) Table 9 shows longer term outcomes for adult TANF leavers and their outcomes over the following three years. Again, employment outcome data are dependent on our ability to identify the adult leaver in the Department of Labor dataset and so does not capture families that moved, were employed by the federal government or were self-employed. The top row shows the total number of leavers for each year from 2009 to 2014. The percentages represent the share of those exiting that were employed in at least one quarter or in all four quarters in the first year after exiting. The wages reported are for leavers who were employed for the entirety of a quarter or all of a year; thus, the table captures the wages that a diligent leaver who keeps his or her job

20 can expect. The second section of the table shows outcomes for leavers in the second year after the exit quarter, and the third section shows outcomes in the third year after exit. Much like Table 8, the basic takeaway is that employment opportunities appear to be increasing and the number of leavers recorded as being employed has gone up over time. In particular, Table 9 shows a marked shift between employment and wages for those in the sample during the 2009-11 period and those in the sample from 2012 onward. Despite these improved circumstances, employment appears to be inconsistent, few were employed for four quarters, and average earnings were below the poverty threshold. For the cohorts in the 2009-11 period, the outcomes are almost indistinguishable across the three-year period aside from average annual wages increasing by 16-17 percent from the first year to the third year after exit. Around 64-65 percent of these leavers were employed in at least one quarter, and 30-31 percent were employed across all quarters in the first year. These numbers are quite similar by the third year from exit, which would be 2011 for the 2009 cohort and 2012 for the 2010 cohort. The 2012 and 2013 cohorts appear to have benefited from an improving economy. In 2012, 69 percent were employed in at least one quarter in the year of exit, rising to 72 percent in 2013. By 2013, 41 percent were employed in all four quarters in the second year after exit, and average annual wages in the first to second quarter went from $10,580 to $11,890 a 12 percent increase. For the 2012 cohort, from the first to third year after exit, average annual wages jumped from $10,273 to $12,969, a 26 percent increase. In addition, 73 percent were employed in at least one quarter of that third year. Note, however, that even in the best case scenario the third year of the 2012 cohort the average earnings were still below the poverty level, even for a family of two. Additionally, the percentage of leavers employed in at least one quarter was quite different from those employed in all four quarters. In the third year of the 2012 cohort, only 43 percent were employed in all four quarters. Tables 10 and 11 are similar but, respectively, show the quarterly employment rates and quarterly earnings results by each cohort of leavers. The tables show the economic shifts at a more granular level. Note that the first quarter column is the quarter after a cohort s exit quarter. Thus, for the cohort that exited in the first quarter of 2009, the first quarter of employment or earnings was the second quarter of 2009. As marked by the shaded boxes in Table 10, by the second quarter of 2013, all but one of the cohorts had quarterly employment rates above 50 percent, and by the second quarter of 2015, seven of the 11 cohorts in the sample had quarterly employment rates over 60 percent. Table 11 shows the wage growth from the first quarter of 2009 through the fourth quarter of 2015 as a cohort progressed over time. Generally, average quarterly wages increase the further a cohort moves from the exit quarter to three years out around 26 percent on average. However, this increase occurs from a low base of roughly $12,000-$14,000 in comparable annual income.

21 Table 10. Quarterly Employment Rates by Post-Exit Quarter (Adult Cases) YEAR AND QUARTER TOTAL QTR 1 QTR 2 QTR 3 QTR 4 QTR 5 QTR 6 QTR 7 QTR 8 QTR 9 QTR 10 QTR 11 QTR 12 20091 1541 50.42 48.6 45.94 42.37 45.68 46.07 44.77 43.54 46.26 46.65 45.03 45.42 20092 1331 51.16 49.06 46.43 50.56 51.08 50.18 49.28 50.11 50.71 49.66 49.28 50.56 20093 1433 49.89 46.96 49.33 49.19 49.61 46.96 47.94 48.29 47.45 47.52 50.31 49.89 20094 1701 44.56 47.61 46.85 47.14 45.56 46.67 47.73 47.2 45.2 47.44 48.26 48.08 20101 1794 48.32 47.04 47.71 45.03 46.04 45.37 43.19 43.08 45.7 44.81 45.31 46.87 20102 1666 51.2 50 46.87 48.67 48.25 47.17 47.29 49.27 48.67 49.81 49.15 52.46 20103 1577 50.53 47.49 49.84 48.95 48 47.24 48.12 49.84 49.65 50.28 53.45 54.85 20104 1677 46.45 46.57 49.31 47.64 46.63 48.65 47.7 47.34 48.59 50.02 50.92 51.87 20111 1933 45.11 44.59 46.14 44.28 46.19 46.45 44.95 45.16 47.9 49.66 50.12 51.06 20112 1584 51.76 50.06 48.92 50.88 50.82 51.38 51.26 52.52 54.29 54.48 54.54 56.06 20113 1730 48.49 46.64 48.78 48.15 48.38 48.67 51.56 52.65 53.69 53.87 55.43 57.39 20114 1960 48.31 49.84 50.2 49.23 50.3 52.14 52.5 53.36 54.48 56.58 55.56 57.9 20121 2015 50.71 50.07 50.22 49.33 52 54.24 54.68 53.64 55.73 56.87 58.36 56.87 20122 1712 51.05 49.24 48.07 50.17 53.03 54.43 53.62 55.95 57.3 58.58 56.25 60.28 20123 1881 51.14 50.93 51.51 52.52 53 53.11 54.81 56.4 57.68 55.44 58.26 59.48 20124 1968 53.15 54.06 55.69 56.4 54.62 58.38 58.94 60.06 59.09 62.85 63.66 64.98 20131 2145 52.21 53 54.35 52.82 53.93 54.77 56.73 55.19 57.85 57.85 59.9 20132 1852 54.42 54.37 54.58 55.83 56.04 57.77 56.15 58.42 60.58 61.87 20133 1800 57.11 54.66 58.44 59.38 58.72 56.72 60.88 61.33 61.44 20134 1825 54.57 56.38 58.02 58.57 57.53 61.26 62.68 64.87 20141 1599 56.22 57.66 57.34 55.9 60.6 61.85 62.72 20142 1443 60.08 59.94 57.58 60.15 61.39 62.37 20143 1379 62.5 58.66 61.49 62.36 63.01 20144 1364 56.52 58.87 61.21 61.36

22 Table 11. Average Quarterly Earnings by Quarterly Leaver Cohort and Post-Exit Quarter* YEAR AND QUARTER QTR 1 QTR 2 QTR 3 QTR 4 QTR 5 QTR 6 QTR 7 QTR 8 QTR 9 QTR 10 QTR 11 QTR 12 CHANGE FROM QTR 1 TO QTR 12 20091 $3,090 $3,249 $3,567 $3,137 $3,476 $3,740 $3,938 $3,463 $3,767 $4,011 $4,067 $3,891 26% 20092 $3,296 $3,562 $3,314 $3,565 $3,762 $3,988 $3,505 $3,909 $4,070 $3,976 $3,978 $4,261 29% 20093 $3,536 $3,292 $3,663 $4,029 $4,130 $3,694 $4,004 $4,269 $4,329 $4,101 $4,283 $4,323 22% 20094 $3,067 $3,445 $3,647 $3,729 $3,494 $3,729 $3,823 $3,882 $3,801 $3,942 $3,930 $4,229 38% 20101 $3,323 $3,575 $3,664 $3,371 $3,633 $3,856 $3,832 $3,786 $3,867 $3,995 $4,097 $3,949 19% 20102 $3,356 $3,619 $3,365 $3,583 $3,765 $3,908 $3,686 $3,838 $3,924 $3,980 $3,878 $3,964 18% 20103 $3,312 $3,159 $3,385 $3,659 $3,655 $3,483 $3,817 $3,781 $4,051 $3,909 $3,928 $3,913 18% 20104 $2,975 $3,379 $3,508 $3,723 $3,696 $3,792 $3,922 $4,142 $4,035 $4,034 $4,056 $4,142 39% 20111 $3,156 $3,379 $3,339 $3,415 $3,413 $3,509 $3,669 $3,511 $3,629 $3,685 $3,900 $3,635 15% 20112 $3,371 $3,379 $3,419 $3,562 $3,702 $3,860 $3,683 $3,825 $3,852 $4,084 $3,834 $4,073 21% 20113 $3,434 $3,325 $3,525 $3,662 $3,874 $3,701 $3,741 $3,925 $4,113 $3,887 $4,166 $4,115 20% 20114 $3,237 $3,379 $3,544 $3,893 $3,504 $3,730 $3,853 $4,098 $3,818 $4,063 $4,227 $4,531 40% 20121 $3,179 $3,347 $3,594 $3,499 $3,615 $3,771 $4,062 $3,874 $4,050 $4,148 $4,430 $4,079 28% 20122 $3,125 $3,456 $3,334 $3,654 $3,633 $3,889 $3,601 $3,852 $3,989 $4,218 $4,067 $4,176 34% 20123 $3,342 $3,158 $3,384 $3,541 $3,794 $3,642 $3,853 $4,033 $4,305 $4,030 $4,078 $4,340 30% 20124 $3,256 $3,419 $3,491 $3,797 $3,709 $3,879 $3,980 $4,313 $4,065 $4,278 $4,395 20131 $3,134 $3,439 $3,506 $3,346 $3,633 $3,709 $3,926 $3,666 $3,954 $4,234 20132 $3,278 $3,426 $3,237 $3,531 $3,768 $4,049 $3,764 $3,958 $3,968 20133 $3,411 $3,413 $3,606 $3,680 $3,992 $3,660 $3,849 $4,075 20134 $3,195 $3,498 $3,574 $3,904 $3,629 $3,913 $4,144 20141 $3,291 $3,356 $3,713 $3,490 $3,670 $3,858 20142 $3,224 $3,642 $3,347 $3,675 $3,848 20143 $3,496 $3,241 $3,458 $3,670 20144 $3,304 $3,639 $3,813 Average 26% *Adult recipient cases (or adult has been sanctioned)

23 Escaping Poverty Table 12. Percentage of Employed Leavers with Annual Earnings in Excess of Federal Poverty Guidelines (Cases by Quarter)* YEAR AFTER EXIT EXIT YEAR COHORT FIRST SECOND THIRD 2009 6006 7.75 9.62 11.32 2010 6714 7.74 9.71 11.58 2011 7207 7.03 9.24 12.36 2012 7576 7.22 10.98 14.16 2013 7622 7.96 11.44 2014 5785 8.79 *Adult recipient cases (or adult has been sanctioned) Table 13. 2011 Leavers' Average Earnings by Number of Quarters Worked and Earnings Above the Federal Poverty Guidelines by Number of Quarters Worked* QUARTERS WORKED TOTAL WORKING, CUMULATIVE QUARTER % SHARE OF TOTAL QUARTERLY WAGES AVERAGE EARNINGS FOR THE COHORT NUMBER EARNING ABOVE POVERTY % ABOVE POVERTY 1 435 7% $1,022 $1,022 2 436 7% $1,303 $2,602 3 425 7% $1,654 $4,961 4 391 7% $1,789 $7,158 6 1.53 5 390 7% $1,993 $9,964 14 3.59 6 420 7% $2,232 $13,391 25 5.95 7 359 6% $2,299 $16,091 22 6.13 8 414 7% $2,511 $20,090 33 7.97 9 442 7% $2,760 $24,830 48 10.86 10 432 7% $3,037 $30,349 63 14.58 11 527 9% $3,650 $40,003 138 26.19 12 1,269 21% $5,554 $64,035 542 42.71 All Leavers 5,940 $2,945 $26,331 891 15.00 *Adult recipient cases (or adult has been sanctioned) As noted above, only a small fraction of TANF leavers appear to graduate out of poverty; however, poverty is defined based on family size and may be affected by the averaging of income numbers. Tables 12 and 13 look explicitly at the percentage of leavers earning in excess of federal poverty guidelines. Note that the sample includes only families recorded in the Department of Labor (DOL) dataset during the first three years after leaving TANF.

24 Table 12 includes adults with no recorded earnings in a particular year. As expected, the table shows that at most around 14 percent of leavers were earning at or above the poverty level by the third year after exit. In the first year, only about 9 percent were above poverty. Interestingly, this number has declined since the 2007 report; in 2002, in the third year after exit as much as 18 percent of the cohort that could be found in the DOL database had annual earning in excess of poverty (DHR 2007, 29). Table 13 shows average earnings and ability to escape poverty for adult leavers over three years by quarter for the 2011 leaver cohort. The table takes all leavers who exited in 2011 and then examines whether they were in the DOL database at some point during the subsequent three years. Some were only in the database for one quarter, some for two and so forth. For those in the database for one quarter, the table then shows the average wage for one quarter; for those who were employed for two quarters, it shows the average wage for two quarters and so forth. By looking at the average wages for that employment cohort, one can determine whether the wages were sufficient to pull the family over the poverty threshold, at least for the period employed. Of the 5,940 leavers who then appear in the DOL database, 1,269 or 21 percent of leavers worked for three years and made an average cumulative wage of $64,035. Of this cohort, 42.7 percent were making above the poverty level for their family, with an average annual wage of about $21,345 (or $64,035 divided by three). In total, about 45 percent had consistent employment for at least two years, as recorded by the DOL. In this snapshot of 2011 leavers who were employed according to the DOL dataset, 891, or 15 percent, at some point made enough to put them above the poverty threshold for their family. This result does not suggest causality, merely association with longevity in the job market and higher wages. Those TANF leavers who were able to sustain employment for three years clearly have better outcomes, but whether this is a result of good luck, good character or program interventions is not clear. Table 14. Factors Associated with Exceeding Poverty Guidelines in the 3rd Year after Exit by Number of Accrued TANF Months at Exit (2009-12)* NUMBER OF ACCRUED TANF MONTHS TOTAL ABOVE POVERTY AVERAGE NUMBER OF DEPENDENT CHILDREN AVERAGE EARNINGS IN 3RD POST-EXIT YEAR EARNINGS IN 3RD POST-EXIT YEAR ABOVE POVERTY GUIDELINES (%) 1-12 Months 24453 3250 1.87 $12,186 13.29 13-24 Months 2467 303 1.95 $12,310 12.28 25-36 Months 340 46 1.91 $13,133 13.52 37-47 Months 105 11 1.79 $13,476 10.47 48+ Months 138 20 1.77 $20,102 14.49 Total 27503 3630 1.88 $12,251 13.19 *Adult recipient cases (or adult has been sanctioned)