I. Introduction

This document presents results from the second stage of the post-time limit tracking study — a survey of former recipients of Temporary Family Assistance (TFA), Connecticut's cash assistance program for needy families with children. The survey targeted individuals from six areas of the state whose TFA benefits were discontinued in late 1997 when they reached the 21-month time limit on cash assistance that was established as part of Connecticut's Jobs First welfare reform initiative. In the first stage of the post-time limit study, individuals who were still not receiving TFA approximately three months after their benefits were discontinued were asked to complete a brief telephone interview describing their current situation and key changes that had taken place since they left cash assistance.¹ In the second stage, individuals who were not receiving TFA were interviewed approximately six months after their benefits were discontinued. During this telephone interview they were asked about their current circumstances and changes in their financial well being and security since their three-month interview.²

The post-time limit tracking study is being conducted by the Manpower Demonstration Research Corporation (MDRC), with funding from the Connecticut Department of Social Services (DSS). MDRC, a nonprofit, nonpartisan organization with more than two decades of experience designing and studying social policy initiatives, is also conducting a full-scale evaluation of Jobs First in two of the six areas that are part of the post-time limit tracking study.

Readers should exercise caution in interpreting the results of the post-time limit tracking study for two main reasons. First, the six-month follow-up period is quite short; much longer follow-up is required to fully understand how families will fare after their welfare benefits are discontinued. Second, although this study explicitly compares families’ situations six months after their TFA had been discontinued with their situations at the three-month point (or in some cases, during their last month on TFA), it is not possible to attribute any changes to the fact that their benefits were discontinued. There is no way to determine how these families would have fared if they had been allowed to continue receiving welfare. MDRC’s full-scale evaluation of Jobs First is comparing the outcomes for Jobs First clients with the outcomes for a randomly selected group of similar clients who are continuing to receive welfare under the prior rules.

It is also important to understand two distinctive features of the Jobs First program. First, owing to the way Jobs First’s time limit is designed and implemented, most of the recipients whose benefits have been discontinued had household income above the welfare payment standard (the maximum grant for their family size) when they reached the time limit. Conversely, most of the recipients with income below the payment standard when they reached the time limit have received extensions of their benefits, and thus are not included in this study.³ Because of this pattern, one would expect that most of the people targeted for this study were employed at the point their benefits were discontinued.

Second, Jobs First includes an unusually generous financial work incentive: All earned income is disregarded (i.e., not counted) in calculating recipients’ cash grants and Food Stamp benefits, as long as their earned income is below the federal poverty level. This policy allows many working recipients to continue receiving their entire cash grant (typically $543/month for a family of three), along with a substantial Food Stamp allotment.

A.  The Jobs First Program and Evaluation

Jobs First was implemented statewide in January 1996, under waivers of federal welfare rules. As noted earlier, the program includes a 21-month time limit on cash assistance receipt and a generous financial work incentive. In addition, TFA recipients are required to participate in employment services targeted to rapid job placement.

1. The Jobs First time limit. Individuals who began receiving TFA after January 1, 1996 entered Jobs First immediately. If they were not exempt from the time limit, their "clock" began with their first month of benefit receipt.⁴ Individuals who were already receiving benefits when the program began were phased into Jobs First gradually during 1996, when they showed up at the welfare office to have their eligibility for benefits recertified; their clocks started with the first month of TFA receipt after that point.

This schedule means that people who entered Jobs First in the first month of implementation (January 1996), received benefits continuously, and were never exempt from the time limit, reached their 20^th month of benefit receipt in September 1997. At that point, they were called in for an "exit interview" to determine whether they qualified for an exemption or a six-month extension of their benefits. In general, extensions are granted to recipients who have made a good faith effort to find employment, but have family income below the welfare payment standard when they reach the time limit, or at any point thereafter. Recipients may receive extensions even if they have not made a good faith effort, if there are circumstances beyond their control that prevent them from working when they reach the time limit.

As noted earlier, statewide DSS data indicate that most of the recipients who show up for an exit interview and have income below the payment standard are receiving extensions. Most of the clients who are not receiving extensions fall into two categories. The majority are people who attended their exit interview and were found to have income that exceeded the payment standard. Others have been denied extensions because they failed to show up for their exit interview. Relatively few clients have been denied extensions because it was determined that they failed to make a good faith effort to find employment.⁵

In effect, this pattern means that a large proportion of the clients who have had their benefits discontinued were employed when they reached the time limit. Because of the large earned income disregard, described above, many of these clients were able to work and receive their entire welfare grant until they reached the time limit. At that point, they experienced a large drop in income because their welfare benefits were discontinued (and their Food Stamp grants may have declined as well). However, they may apply for an extension later if their income drops.

2. The Jobs First evaluation. MDRC is conducting a multi-faceted evaluation of Jobs First in New Haven and Manchester, under a contract with DSS. The study uses a random assignment research design, in which about 6,000 welfare applicants and recipients were assigned to one of two groups: the Jobs First group, whose members are subject to all of the rules described above, or the AFDC group, whose members are subject to the prior welfare rules (i.e., they have no time limit and also do not receive the enhanced earned income disregard). MDRC is studying the members of these two groups during a follow-up period of several years; any differences that emerge between the groups’ members will be attributable to Jobs First. These differences are known as the program’s impacts.

This post-time limit tracking study complements the Jobs First evaluation, but is not directly part of it (indeed the post-time limit study is occurring in four areas that are not part of the full-scale evaluation). As discussed below, the post-time limit study is intended to describe the post-welfare circumstances of clients who reach the time limit and do not receive extensions. However, the study cannot measure the impact of Jobs First because there is no comparison group against whom to compare the clients who are being surveyed. The first evidence on the impacts of Jobs First — that is, the differences between the Jobs First and AFDC groups — will appear in an interim evaluation report scheduled for 1999.

B. The Post-Time Limit Tracking Study

As discussed above, the post-time limit tracking study focuses on the early experiences of a sample of cases who were the first to reach their 21-month TFA time limit and not receive a cash aid extension.

To date, not much is known about the economic status and household stability of families who are no longer receiving welfare. Most existing research has focused on the financial circumstances of women as they cycled on and off Aid to Families with Dependent Children (AFDC). However, the employment and economic experiences of women who chose to leave welfare could be different from those whose welfare benefits were discontinued. The findings presented in this document seek to fill in some of the gaps, and provide a snapshot of the lives of several hundred families at very early stages of their post-welfare period.

1. The sampling frame. Cases eligible to be surveyed for this study include those that reached the 21-month TFA benefit time limit and did not receive an extension. These cases were selected from six DSS district offices: Bridgeport, Hartford, Manchester, New Haven, Norwich, and Waterbury.

Figure 1 illustrates how the survey sample was selected. To identify the sample, DSS provided MDRC with computerized files containing the names of all recipients who were scheduled for a 20-month exit interview in September or October 1997 (the first months in which such interviews took place), along with information on the outcome of the exit interview for some of the cases. MDRC began by eliminating cases from the non-study sites and cases that were coded as having received an extension during the exit interview. The remaining cases were looked up on the Eligibility Management System (EMS), and those that were found to be receiving TFA just before the survey began were also eliminated from the sample (these cases had either been granted extensions initially but had not been coded as such, or had been denied extensions initially but had returned to welfare in the meantime). The remaining cases constituted the survey sample; however, respondents who were found to be receiving TFA when contacted were not asked to complete the interview.

As shown in Figure 1, there were 1,644 cases in the study sites scheduled for exit interviews in September or October 1997 (773 and 871 cases, respectively). Of these, 277 cases in the September cohort were dropped because the DSS file indicated that they had received an extension during their exit interview. From the remaining 496 cases in the September cohort, another 117 were dropped because they were receiving TFA in December 1997, and 42 cases were dropped because it was determined that they had not used up all 21 months of their time limit clock.⁶ Ten cases were used for a pre-test of the survey instrument, leaving a total of 327. Nearly 500 cases in the October cohort were dropped because the DSS file indicated that they had received an extension during their exit interview. From the remaining 398 cases in the October cohort, 266 cases that fit the criteria described above were selected.⁷ Thus, the full sample for the study includes 593 cases (327 from the September cohort and 266 from the October cohort).

2. Survey method. The post-time limit survey was conducted by the Response Analysis Corporation (RAC), under a subcontract with MDRC. RAC used a Computer Assisted Telephone Interviewing (CATI) technique.⁸ To encourage survey participation, respondents were offered a $10 incentive to complete the survey. When RAC’s phone center was unable to reach a sample member by telephone, field trackers were sent to their last known address to encourage them to complete an interview. Interviewing for the three-month survey effort began in the second week of January and concluded in the second week of April 1998. Interviewing for the six-month interview began in the first week of April and concluded in the last week of July 1998.

3. The contents of the interview. The post-time limit survey was designed to elicit information from respondents regarding their income, employment, household composition, financial well-being, and food sufficiency experiences six months after their TFA benefits were discontinued as well as indicate changes in these areas since their three-month follow-up interview. On average, respondents were interviewed about 7 months after their TFA grant was discontinued.⁹ The average length of the interview was about 15 minutes. The survey was administered in English and Spanish.

4. Survey response rates and representativeness. Of the 593 sample members eligible to be surveyed, 480 or 82 percent¹⁰ were contacted at the six-month point. However, because 32 of the 480 respondents were receiving TFA at the time of the interview, only 448 sample members were asked to complete the survey (See Figure 1).

An analysis was performed to determine whether estimates based on survey data are systematically biased by the absence of completed interviews for some sample members. The analysis showed there were some differences in the characteristics of responders and non-responders, but these differences were not systematic. See Appendix A for an in depth discussion of the response analysis.

It is also important to note that the post-time limit tracking study sites were not randomly selected from all the sites or district offices in Connecticut. Consequently, caution should be exercised when generalizing these survey findings. Nevertheless, about three-fourths of the statewide TFA caseload is represented by the districts in this study.

5. Characteristics of survey respondents. As shown in Table 1, the largest proportion of respondents (24 percent) was from New Haven, followed by Hartford (19 percent), Bridgeport (18 percent), Norwich and Waterbury (15 percent), and Manchester (11 percent).

The vast majority of the respondents are female (95 percent). The average age of respondents at the time of the interview was 34. About 49 percent of the respondents were between 25-34 years old, and 34 percent were between 35-44. Roughly 11 percent of the respondents were between 20-24. Six percent were 45 and older.¹¹

The total sample is comprised of three major ethnic groups that are nearly equally represented. Thirty-five percent of the survey respondents are white non-Hispanic, 33 percent are Hispanic, and 31 percent are black non-Hispanic.¹²

More than half of the respondents (54 percent) have a high school diploma or GED and about 7 percent went to college or have a college degree. A substantial number of respondents (40 percent), however, did not graduate from high school or obtain a GED.

6. Demographic differences by district office. As shown in Appendix Table B.1, the ethnic breakdown of the sample is significantly different across the sites. As one might expect, there are significantly more black and Hispanic than white respondents in the urban sites. In Bridgeport, 42 percent of the respondents are black, 48 percent are Hispanic and 8 percent are white. The ethnic composition of the Hartford site is similar: 33 percent of the respondents are black, 57 percent are Hispanic, and 11 percent are white. In contrast, the Norwich, Manchester and Waterbury samples are comprised largely of white respondents. The New Haven site was composed of more black respondents (41 percent), than white (30 percent), or Hispanic respondents (27 percent).

II. Survey Results

Results are presented below for each of the five main topics covered in the survey: employment, housing, income, financial security, and food sufficiency. For each topic, the discussion begins by describing the status of the 448 sample members at the time of the six-month interview (or, in some cases, in the month prior to the interview). Next, the discussion focuses on key changes that have occurred since the three-month follow-up interview for 373 sample members who were interviewed at both points. Consequently, changes in employment, total household income, and food sufficiency were determined by comparing the six-month responses of these 373 sample members with their three-month responses. ¹³ The responses of the sample members who were interviewed at both points were also used to indicate changes in housing, household composition, and employment since their last benefit month.¹⁴ Finally, findings for the individual district offices are discussed in most instances when significant differences exist.

The tables are organized in a similar manner. In most tables, the top panel shows the status as of (or just before) the six-month follow-up interview of 448 respondents, while the bottom panel describes changes that have occurred since the three-month interview for the 373 sample members who were interviewed at both points. Tables 1 to 7 include results for the full sample, and appendix Tables B.1 to B.7 show the results by district.

As previously mentioned, it is important to recognize that sample members who had returned to welfare and were receiving TFA when contacted for a three or six-month interview were not surveyed. Consequently, we do not know what changes in their circumstances caused them to reapply for TFA. We do know, however, that there were fewer respondents receiving TFA at the six-month follow-up (N= 32 ) than at the three-month follow-up point (N = 42). Also, of the 42 respondents who were receiving TFA at the three-month interview point, 35 were contacted again at the six-month follow-up. Fifteen of these 35 respondents reported they were still receiving TFA.

A. Employment, Earnings and Job Characteristics

The survey asked respondents about their current employment status, and focused a number of detailed questions on the characteristics of their primary current job (generally the job in which they work the most hours). The survey also asked less detailed questions about other jobs the respondents had held since their three-month follow-up interview.

1.  Employment status six months after TFA benefits were discontinued. As shown in the top panel of Table 2, approximately 83 percent of the respondents were employed at the time of their six-month interview, and 6 percent reported working at more than one job.

The next section of Table 2 focuses on respondents who were employed at the time of the interview. It shows that, on average, employed respondents worked 35 hours per week at all jobs. The average earnings per week from all jobs were $270. (Table 3 provides additional information on the primary job held by employed respondents. For example, Table 3 shows that respondents worked an average of 34 hours per week in their primary job and earned, on average, $7.82 per hour.)

About 17 percent of the respondents were not employed at the time of their six-month interview. When asked why, 48 percent indicated they "could not find work," 23 percent reported they were "taking care of someone," 10 percent indicated they had "a health problem or were pregnant," 7 percent said they were "in school," 7 percent had transportation problems and 3 percent provided some other reason.¹⁵ Fifty-five percent of the respondents who were not employed indicated they were currently looking for work.

2. Changes since last benefit month and three-month interview. The bottom panel of Table 2 compares respondents’ employment status in their last benefit month, and at their three-month and six-month interviews. These findings were derived from an analysis of the responses of 373 respondents who completed a three-month and a six-month survey. The overall employment rate was very similar at all three points, ranging from 82 to 85 percent (all results not shown in a table). Similarly the employment status of most of the respondents’ was the same at all three points. About 72 percent were employed at all three points. Conversely, 7 percent of the respondents were not employed during their last benefit month nor at their three or six-month interviews (See bottom panel Table 2).

The bottom panel of Table 2 also shows that 78 percent of the respondents were employed at both the three and six-month follow-up points. A number of respondents (6 percent) were not working at the time of their three-month interview but had found employment by the six-month follow-up point. Conversely, 5 percent were employed at the three-month point, but not at the six-month point. Of those who were working at the time of their three-month interview, 92 percent were still employed at the same job three months later.

Further analysis (results not shown in a table) found that 33 percent of the respondents who were employed at both the three-month and six-month points were working more hours at the time of their six-month follow-up point than they were at their three-month follow-up point. Conversely, 23 percent were working fewer hours. This analysis also found that the respondents who were working at the six-month point, but not during the month of their three-month interview, were working an average of 35.2 hours per week, and earning, on average, $6.29 per hour.

3. Employment differences by district office. The percentage of respondents who reported they were currently employed at the time of their six-month interview differed significantly across the sites. For example, current employment ranged from as low as 76 percent in Waterbury to as high as 92 percent in Manchester. Bridgeport respondents had the next highest rate of current employment (91 percent), followed by Norwich respondents (85 percent), Hartford (81 percent) then New Haven (78 percent) respondents (See Table B.2).

There were also significant differences across sites with respect to the number of hours worked per week by respondents at their primary job. For example, although the Manchester site had the highest percentage of respondents employed at the six-month interview point, those respondents reported working the least number of hours per week (30 hours per week) at their primary job. Hartford respondents reported working the most hours per week (36 hours per week), followed by Bridgeport, New Haven and Waterbury (34 hours per week) then Norwich respondents (33 hours per week) (See Table B.3).

B. Household Composition and Housing Arrangements

Some have suggested that discontinuing welfare benefits could lead to changes in household composition. That is, former recipients might move in with friends or family to cut costs. Families could also send children to live with relatives.

1. Household composition six months after TFA benefits were discontinued. The top panel of Table 4 describes respondents’ housing status and household composition at the point of the six-month interview. The most common household (31 percent) consisted of three persons. There were also some fairly large households: 13 percent of the households contained five persons, and 14 percent contained six or more persons. Just about all (97 percent) of the households included one or more children. The absence of children in 3 percent of households is attributed to persons moving out of the respondent’s household since their benefits were discontinued. About 43 percent of the households included at least one other adult beside the respondent.

The Census Bureau defines a unit as crowded if it has more than one person per room. As indicated in Table 4, 11 percent of the respondents live in crowded conditions by this definition.

2. Changes since last benefit month and the three-month follow-up interview. The bottom panel of Table 4 shows changes that occurred since the last benefit month for the 373 respondents who completed both a three and six-month survey. Overall, there were few major changes in respondents' living arrangements. For example, approximately 24 percent of the respondents moved at some point between their last benefit month and six-month interview. Of the respondents who moved, about 11 percent (2 percent of all respondents) indicated that one or more of these moves led to a home that was worse than their previous home (See Table 4). Seven respondents (2 percent) reported being homeless at some point between their last benefit month and their six-month interview. However, at least 2 of the 7 respondents also reported being homeless in the year prior to the discontinuance of their TFA benefits.¹⁶

An additional analysis (results not shown in a table) also showed there was no change in household size since the three-month interview for 71 percent of respondents. Fifteen percent reported an increase in household size and 14 percent reported a decrease.

3. Results by district office. The analysis of the number of rooms in the respondents’ home indicated statistically significant differences across sites. However, the actual numbers of rooms in the living quarters of these respondents were nearly the same in all sites. (See Table B.4.) Hartford respondents reported the least number of rooms in their home (4.6 rooms) and Norwich and Manchester respondents reported the most number of rooms in their homes (5.2 rooms). Waterbury respondents reported the next highest number of rooms in their home (5.1 rooms), followed by Bridgeport respondents (5.0 rooms) then New Haven respondents (4.9 rooms).

C. Household Income

For this analysis, total household income was calculated for the month prior to the six-month interview by asking respondents a series of questions about income from various sources. The survey asked about income for all household members. However, income from household members other than the respondent was counted in the total only if the respondent indicated that this income helped to support her and her children. Specifically, 26 percent of respondents indicated that other members of their household had income. However, of these, over one third (37 percent) indicated that the other household member’s income did not help to support them or their children; thus, this income was excluded in the total household income calculation.

1. Status in the month prior to the six-month interview. As shown in the top panel of Table 5, 14 percent of the respondents reported their household income was between $1-$499, 29 percent reported income between $500-$999, and 32 percent reported income between $1,000-$1,499. The average total household income in the month prior to the six-month interview was $1100. The average income for respondents who reported that their household was comprised of three individuals (31 percent of respondents) was $1046. For respondents with a household size of four (23 percent of respondents), the average income was $1141. (Income by family size is not shown in the table.)

Several separate analyses were conducted to determine how respondents sustained themselves without employment, or on little or no income in the month prior to the interview.

For example, 3.2 percent of the respondents reported no household income. An analysis showed 9 of these 14 respondents reported borrowing money from friends or family to sustain themselves during the month prior to the interview. One of the 14 respondents also reported that they lived with family and did not pay rent; consequently their family members may have provided more than rent-free housing. However, none of the 14 respondents indicated yes to the "does the income received by other members of your household help to support you or your children" survey item.

An analysis of the 14 percent of respondents that reported household incomes between $1-$499 was also conducted. The average household size for this group was 3.7, and more than half (56 percent) of the respondents in this category did not have another adult in their household besides themselves. Nearly half (48 percent) reported they were employed at the six-month follow-up. Finally, within the sites, the proportion of cases in this income category ranged from 10 percent in Manchester to 18 percent in New Haven.

Another group of interest is the 57 respondents (13 percent of the sample) that reported no one in their household was employed. Eight of these 57 respondents (14 percent) reported someone in the household received income from other sources. Thirty-five of the remaining 49 respondents reported using some of the following strategies to cope: borrowing money, borrowing food, or dipping into savings to support their families during the month prior to the interview.

Another comparison of interest (results not shown in a table) is how respondents’ total income post-TFA measures up against what they previously could have received from TFA and Food Stamps (without supplemental income from earnings).¹⁷ About two-thirds of the respondents (between 65-67 percent¹⁸) had higher incomes six-months after their TFA benefits were discontinued than they could have received from TFA and Food Stamps (without working).¹⁹ About a third (between 33-35 percent²⁰) of the respondents had less income than they could have received from TFA and food stamps. On average, respondents reported that they had received $492 in TFA cash assistance in their last benefit month.

An analysis of the 35 percent of respondents with less income in the month prior to their interview than they could have received in public assistance was conducted.²¹ Thirty-nine percent of the respondents in this group had income levels between $1-$499 at the time of the interview. Sixty-three percent of these respondents reported that they were employed at time of their six-month interview.

With the exception of respondents from the New Haven (41 percent) site, approximately one third of all respondents in Norwich (37 percent), Waterbury, (33 percent), Bridgeport (30 percent), and Manchester (29 percent) reported incomes lower than the maximum they could have received from TFA and Food Stamps.²²

2. Changes since the three-month follow-up interview. The bottom panel of Table 5 shows the results of a direct comparison of income in the month prior to the three-month interview and the month prior to the six-month interview for the 373 respondents who were interviewed at both points. A little more than half of these respondents (51 percent) reported higher household income at the six-month follow-up point, than at their three-month equivalent. Roughly 41 percent of these respondents experienced a decrease in income by their six-month follow-up point and the income of 5 percent of these respondents remained about the same at both points.²³ As indicated in the bottom panel of Table 5, 21 percent of these respondents experienced income increases of $1-299, and 30 percent had income increases that were $300 or more. Twenty-one percent of these respondents had a decrease in income that was $1-299, and 20 percent experienced decreases in income that were $300 or more.

An additional analysis (results not shown in a table) indicated that the average total household income for these 373 respondents changed from $991 at the three-month point to $1100 at the six-month point.

3. Results by district office. There were significant differences between sites with regard to the amount of income received from child support reported by respondents. Hartford respondents reported the least amount of income from child support ($105) and Manchester respondents reported the most ($313), followed by New Haven ($257), Bridgeport ($245), Waterbury ($244) then Norwich respondents ($218).

D. Financial Security and Other Measures of Well-Being

Research has shown poverty has harmful effects on families, and especially on children, in part because it implies less access to basic necessities. This section examines respondents' access to basic necessities such as food and medical care.

1. Status in the month prior to the six-month interview. Recipients who leave welfare for work are eligible for transitional Medicaid coverage for two years. As shown in Table 6, seven percent of the respondents (N = 29) indicated they were not covered by health insurance six months after their TFA benefit had been discontinued. Twenty-one of these 29 respondents were working at the time of their six-month interview. Table 6 also indicates that 6 percent of all respondents (N = 28) reported there were children in their household that were not covered by medical insurance. Twenty-one of these 28 respondents reported they were employed at the time of their six month interview.

Respondents reported using various strategies to make ends meet in the month prior to their six-month interview. Approximately 61 percent indicated they delayed paying their bills in the month prior to their interview. Thirty-seven percent reported borrowing money from family or friends and 27 percent reported borrowing food. Thirteen percent of the respondents indicated they got food from a church, soup kitchen or food bank during this period.

2. Changes since the three-month interview. Respondents who indicated that they had used a particular strategy to make ends meet in the month prior to the six-month interview were then asked whether they had used that strategy more, less, or the same amount as in the month before their three-month interview. For example (results not shown in a table), of those who indicated that they had delayed paying bills in the month prior to the six-month interview, 38 percent indicated that they did so more in the month before their six-month interview than in the month before their three-month interview.

Approximately 48 percent of the respondents reported they spent less money in the month prior to their six-month interview than in the month prior to the three-month follow-up. More than a quarter of the sample (26 percent) also reported saving less money and about 21 percent reported they worked more hours. (See Table 6, bottom panel) Table 6 also shows that 58 percent of the respondents bought smaller or less expensive meals in the month prior to their six-month interview than they did at the three-month follow-up point.

There were, however, a number of respondents who at the time of the six-month interview, were not using coping strategies that they had used at the three-month point. An analysis of the responses of the 373 respondents interviewed at both points in time showed that fewer of these respondents delayed paying their bills, borrowed food or money from friends or family, or got food from a food bank by their six-month interview. Specifically, the proportion of respondents who reported they delayed paying bills went from 66 percent at the three-month point to 60 percent at the six-month point; the percentage of respondents who borrowed food went from 30 percent at the three-month point to 26 percent at the six-month point; the percentage of respondents who reported borrowing money dropped from 39 percent to 36 percent; and the percentage of respondents who reported getting food from a food bank decreased from 15 percent to 12 percent.²⁴

When asked about their standard of living (e.g., their food, housing, medical care, and recreation) 34 percent of respondents indicated they were less satisfied with their standard of living in the month prior to their six-month interview than they were at their three-month interview. Forty-six percent indicated they felt about the same, and 20 percent of the respondents reported they were more satisfied. This pattern is somewhat surprising in light of earlier findings, which indicated that about half the respondents reported higher household incomes at the six-month point than at the three-month point.

E. Food Sufficiency

The food sufficiency questions used in this survey are a subset of items used by the United States Department of Agriculture’s Food and Consumer Service in establishing national food-security benchmarks.

1. Food Sufficiency status six-months after TFA benefits were discontinued. As shown in Table 7, the majority of respondents reported that their families either "always" had enough to eat (40 percent) or had "enough," but not always the kinds of food they wanted (45 percent) at the time of the six-month interview. Twelve percent reported their families "sometimes" did not have enough to eat, and 3 percent of the respondents reported this was "often" the case.

An additional analysis (not shown in a table) indicated that 22 percent of the respondents who reported food insufficiency reported monthly income of less than $500. Approximately 53 percent of the food insufficiency respondents also did not have another adult in their household and a quarter of them reported that their households consisted of 5-6 people. Seventy-seven percent of these respondents reported they were employed at the time of their six-month follow-up.

A number of the respondents also reported they "often" (21 percent) or "sometimes" (38 percent) relied on low cost foods to feed their children because they ran out of money to buy food. Approximately 40 percent of the respondents reported this was "never" the case.

2. Changes since the last benefit month and three-month interview. The analysis of the responses of 373 respondents with data from both the three and six-month follow-up interviews suggests that the respondents experienced changes in food sufficiency between the respondents last benefit month and their six-month interview. As indicated in the bottom panel of Table 7, when respondents were asked how much they relied on low cost food to feed their children during their last benefit month: 14 percent said "often," 30 percent said "sometimes" and 57 percent said "never." However, at the six-month interview 21 percent said "often," 38 percent said "sometimes" and 41 percent said "never." As noted earlier, however, it is not possible to attribute this apparent change to the fact that benefits were discontinued.

3. Results by district offices. Within the sites, the percentage of respondents who reported they "sometimes relied on low cost food to feed their children" ranged from 24 percent in Norwich to 51 percent in New Haven (see Table B.7).

Notes:

¹As discussed below, under certain circumstances, individuals may receive extensions of their cash assistance at some point after their benefits are discontinued.

²Some people who were interviewed at the six-month point were not interviewed at three-month point, either because they could not be located, refused to complete an interview, or they were receiving TFA. These individuals obviously could not be asked about changes since their three-month interview.

³As discussed further below, some recipients had their benefits discontinued because they did not attend an "exit interview" in their 20^th month of assistance; their income situation was likely unknown to DSS.

⁴Families are exempt from the time limit if no adults in the household are required to participate in employment-related activities.

⁵This pattern may change in the future because clients who are granted extensions will have their benefits discontinued if they fail to cooperate with employment-related mandates during the extension period.

⁶The 117 cases probably include some who were granted an extension at the exit interview (even though they were not coded as such on the files received from DSS) and others who were initially denied an extension but had returned to TFA by December.

⁷nly 266 additional cases were needed to complete the fielded survey sample goal. They were selected from cases scheduled for an exit interview in October, proportional to their representation by site in the September cohort. Since not all the cases in the October cohort were needed for the fielded sample, the cases were sorted in random order so that each case within a site had an equal chance of being looked up on the EMS.

8With CATI, the survey instrument is imbedded in a computer program that displays the questions in the appropriate order and allows the interviewer to type in the responses. The data are directly entered into the program, which does not accept inconsistent responses, resulting in a cleaner data file.

⁹The telephone numbers extracted from the EMS for a substantial number of sample members were no longer valid by the time RAC attempted to contact them. Consequently, the field period was extended so that field trackers could be sent to the last known address of these sample members to encourage them to participate in the study.

¹⁰Five of the 593 cases were removed from the base for calculating the response rate because the respondent did not speak English or Spanish.

¹¹The demographic data for respondents were extracted from the EMS. Respondents were not asked to verify its accuracy.

¹²The ethnicity field for 1.3 percent of the respondents was blank on the EMS.

¹³Respondents who did not complete a three-month survey were asked to compare their current levels of food sufficiency with levels three months before their six-month interview. Questions regarding their experiences with the Jobs First extension process were also included in their interview. Although their food insufficiency and extension process experience data were processed they were not included in this report. The responses, however, of the respondents who did not complete a three-month interview are similar to the respondents who did complete a three-month survey. (See Connecticut Post-Time-Limit Tracking Study: Three-Month Survey Results, by Jo Anna Hunter-Manns, et. al. Tables 7, 8, 9.)

¹⁴The three-month interview included a number of questions about respondents’ status in their last month of TFA receipt.

¹⁵These data were missing for 3.2 percent of these sample members.

¹⁶The housing status of two of the seven respondents in the year prior to their last benefit month is unknown because these data were missing.

¹⁷This comparison is based on maximum TFA and Food Stamps benefits for the reported household size at the time of the interview. Note that this may be different from the respondent’s household size while on TFA, due to changes in household composition since that time, and also because the household may include individuals who would not be considered eligible members of the respondent’s TFA or Food Stamps case.

¹⁸A range was provided because when these calculations were performed excluding the income of "other" household members that, according to the respondent did not financially assist them, 65.3 percent of the households were above the combined TFA and FS total for their family size. However, when the income of "other" household members was included in these calculations, 67.1 percent of the households exceeded the combined TFA and FS total for their family size.

¹⁹A family of three with no other income could receive $784 in benefits per month — $543 in TFA, and $241 in Food Stamps. For a family of four without earnings, maximum TFA and Food Stamps benefit amounts are $639 and $283, respectively, for a total of $922.

²⁰A range was provided because when these calculations were performed excluding the income of "other" household members that, according to the respondent, did not financially assist them, 34.7 percent of the households were below the combined TFA and FS total for their family size. However, when the income of "other" household members was included in these calculations, 32.9 percent of the households were below the combined TFA and FS total for their family size.

²¹The total household income measure which excludes the income of other household members that did not financially assist the respondent was used for this analysis.

²²The total household income measure which excludes the income of other household members that did not financially assist the respondent was used for this analysis.

²³These data were missing for 2.4 percent of the sample.

²⁴These six-month interview figures do not exactly correspond to the figures in the top panel of Table 6. This is because the figures presented here include only the 373 respondents who were interviewed at both points in time.