The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PROWRA) stipulates that able-bodied, childless adults may only receive food stamps for 3 months in a 36-month period unless they work at least 20 hours a week; participate in an approved work or training program; or live in an area that has been waived from the time limit due to either an unemployment rate higher than 10 percent, or insufficient jobs. A provision of the Balanced Budget Act of 1997 permits states to exempt up to 15 percent of their unwaived, unemployed, childless able-bodied (ABAWD) population from the PRWORA three-month time limit. Since this population is at risk of losing their food stamps without appropriate work opportunities, these exemptions give states broad discretion to allow selected ABAWDs to continue receiving food stamps beyond their initial three-month period of eligibility. Because the exemptions are limited to 15 percent of each state’s unwaived unemployed ABAWD caseload, states must exercise this discretion responsibly in order to grant the maximum number of exemptions without exceeding their average monthly allotment.

The U.S. Department of Agriculture Food and Nutrition Service (FNS) has prepared this guidance to assist states in managing their exemptions by estimating the effects of certain policy choices states may make in deciding which ABAWDS are in the greatest need of exemptions, and how long those exemptions should last. This guidance gives states background information on how the exemptions were determined, discusses the food stamp participation dynamics of the ABAWD population, describes the method FNS has created to help states determine the number exemptions it can use to exempt a certain number of ABAWDs for various periods of time, and provides information on specific groups states may want to exempt.

FNS would like to stress that this is guidance only and is intended to assist states to set policy and to be responsible for oversight of implementation of these policies. Much of the information is derived from the Quality Control (QC) sample of the entire food stamp caseload, and not on an actual sample of the ABAWD caseload. Bear in mind that although this information is the best available at this time, ABAWD caseloads are likely to have changed dramatically since the passage of welfare reform and it is difficult to say with any certainty how closely the existing data reflect current conditions. Finally, FNS would like to stress that there are no penalties if the policies the state agencies use or develop result in states using more than their allocated exemptions in a given year, as the law provides for an adjustment in the exemptions that will be allowed in subsequent years.

###### Arriving at the By-State Numbers of Exemptions

The Balanced Budget Act required that the average number of exemptions be no more than 15 percent of the number of "covered individuals" per state. "Covered individuals" is the statutory term for ABAWDS. They are those individuals who are subject to the time limit and are not already fulfilling the work requirements, living in an area that has been waived from the time limit, or currently in their first or second three months of eligibility. The Balanced Budget Act also specified that for FY 1998 the number of covered individuals would be based on FY 1996 QC data, adjusted as necessary.

To arrive at the number of covered individuals FNS began with the entire FY 96 QC data file, and then made adjustments by:

- excluding recipients exempted from the ABAWD provisions
- excluding to the extent possible those non-citizens made ineligible for food stamps after August 22, 1997
- excluding the number of recipients who were complying with the work requirements
- excluding to the extent possible those people who were at the time in their initial first three months of eligibility
- adjusting this data to reflect the actual change in each state’s caseload between FY 96 and FY 97 and the estimated national caseload change between FY 97 and FY 98, and
- excluding those individuals living in waived areas.

FNS then multiplied the number of covered individuals by 15 percent to arrive at the number of covered individuals for each state.

Based on this methodology, FNS has authorized approximately 64,000 average monthly exemptions for ABAWDs nationwide and made allocations from this total to the states. FNS would like to stress that the average number of exemptions allocated to each state for this fiscal year is based on the number of covered individuals in FY 96 (before the ABAWD time limits took effect) and therefore will likely be greater than 15 percent of the current number of covered individuals in areas that have implemented the time limits.

The average monthly exemptions are allocated on a case-month basis, so that the annual authorized number of case months for each state is 12 times the average month. If this level of exemptions is not used by the end of the fiscal year, the state may carry over the balance. If more exemptions are used than authorized in a fiscal year, the state’s allocation for the next year will be reduced.

###### Taking into Account the Participation Dynamics of the ABAWD Population

The ABAWD population is dynamic - its members will cycle on and off the Program, or into and out of ABAWD status for a variety of reasons. FNS has analyzed longitudinal data from the Survey of Income and Program Participation (SIPP) and determined that the average ABAWD who reaches the 3-month time limit would likely remain on the program an additional 2.66 months if granted an exemption. Therefore, state exemption policies that exempt a number of persons in one month will need to account for the ongoing participation of these people in subsequent months.

This is because the average ABAWD is liable to come on and off the program several times in a year, and may only need an exemption for a period of less than 3 months at any given time. Thus, it is necessary for state agencies to account for these participation dynamics in calculating how many exemptions it can distribute in any given month. For example, assume a state agency has been allocated an average of 300 exemptions a month. In October, the state agency grants exemptions to the 300 ABAWDs who reach the time limit in that month and become eligible for an exemption. These ABAWDs will, on average, need that exemption for October, November and December. So, when November rolls around and another 300 ABAWDs become eligible for an exemption, the state agency has already used up its average monthly number of exemptions and on average, won’t be able to grant new ones until the first group of ABAWDs leave the Program by the end of December.

The question arises then, "How can a state figure out how many exemptions it can maintain (not grant, but maintain) in any given month without exceeding the allocated number of exemptions?"

###### The Method

FNS analyzed longitudinal data and determined that ABAWDs would generally leave the Program on their own accord within 2.66 months after reaching the time limit. In other words, if a state wanted to grant permanent exemptions to all ABAWDs as they become eligible for an exemption for an indefinite period of time, for every 100 ABAWDs, a state would actually need to use 266 case months worth of exemptions.

FNS also determined how many case-months worth of exemptions a state would have to use if it placed 3- and 6-month time limits on the exemptions.

- 266 case-months worth of exemptions to serve 100 ABAWDs in an average month if NO TIME LIMIT were placed on the exemptions
- 206 case-months worth of exemptions to serve 100 ABAWDs in an average month if a 6-MONTH TIME LIMIT were placed on the exemptions
- 141 case-months worth of exemptions to serve 100 ABAWDs in an average month if a 3-MONTH TIME LIMIT were placed on the exemptions.

Knowing the participation dynamics of the ABAWD population, and given the average number of monthly exemptions the state is allocated, how would a state figure out how many of those exemptions it could use in any given month? How does the information above help states determine how many exemptions it can use in any given month and not exceed their allocations?

To take into account this participation dynamic, the state could take a portion of the average monthly exemptions and grant fewer new exemptions each month, but continue exemptions granted in prior months while staying within the state exemption total. For example, assume that a state has 266 average monthly exemptions. Using the information from above, that each new exemption, on average, lasts 2.66 months, suppose that in October, the state agency [only] grants exemptions to 100 ABAWDs (266 exemptions divided by 2.66). They use 100 exemptions in October, and on average 166 more in subsequent months. For purposes of this example, assume that all 100 use an exemption in November and 66 use an exemption in December, and that all of them will leave on their own accord by January. However, in November, the state grants exemptions to an additional 100 ABAWDs. So, for the month of November, the state uses 200 exemptions. This second group of ABAWDs first granted exemption in November remain exempt for November and December and 66 persons will use an exemption in January before leaving on their own accord in February. In December, the state grants exemptions to an additional 100 ABAWDs. So, for the month of December, the state has used 266 exemptions (100 first granted in December, 100 from November, and 66 from October). In January all the first group has left , but new ABAWDs become eligible. If the state maintains this rate of granting exemptions, the state can anticipate using 266 exemptions a month.

The following illustrates the ways that a state can estimate its monthly allotment of exemptions under different time-limit and start-date scenarios, and how many ABAWDs can be served via those exemptions in a given month. These examples are for the state of Alabama, whose average of 1,650 case-months worth of exemptions represents approximately 15 percent of its unwaived able-bodied caseload:

###### Table A - No time limit on the exemptions

Table A presents for each state the original allocation of average monthly exemptions for the 12-month period that encompasses FY 1998. The table then calculates the number of exemptions that each state could issue each month depending on when the state chooses to issue the exemption. The table further assumes that the state places no time limit on the exemptions and that the ABAWD population exhibits the normal participation dynamic.

For example, Alabama has been allocated an average of 1,650 exemptions a month. If Alabama started using the exemptions in October and anticipates using them for the entire year, it would divide 1,650 exemptions by 2.66 months to determine how many exemptions it could use in any given month and not exceed its allotment (1,650/2.66 = 620). The state could exempt 620 new ABAWDs a month and not exceed its allotment. It would have sufficient exemptions to continue to exempt these persons in their ensuing months of program participation.

If Alabama started using the exemptions in January and only had 9 months left to use them, it would multiply the number of originally allocated average monthly exemptions by 12 to determine the yearly amount, then divide that number by the remaining 9 months to determine the average monthly number of exemptions the state has for the last 9 months of the year. Then it would take this number and divide by 2.66 months to determine how many exemptions it could use each month from January through September and not exceed their allotment ((1,650 x 12)/9 = 2,200)); (2,200/2.66 = 827). Alabama could exempt 827 ABAWDs each month from January through September without exceeding its allotment.

If Alabama started using the exemptions in April and only had 6 months left to use them, it would multiply the number of originally allocated average monthly exemptions by 12 to determine the yearly amount, then divide that number by the remaining 6 months to determine the average monthly number of exemptions the state can use over the last 6 months of the year. Then it would divide this number by the 2.66 months to determine how many exemptions it could use each month from April through September and not exceed its allotment ((1,650x12/6) = 3,300)); (3,300/2.66=1,241). Alabama could exempt approximately 1,241 ABAWDs each month from April through September without exceeding its allotment.

###### Table B - 6 month time limit

Table B is similar to Table A. The key difference is that the calculations in Table B assume that states limit the length of the exemption to 6 months, whereas in Table A, there was no limit. State agencies can use the same method if they want to determine how many case-months worth of exemptions it would need if it put a 6-month time limit on the exemption. However, as explained earlier, for a 6-month time limit it would need to reserve 2.06 months for each exemption granted. If Alabama implemented the provision in October, it has 12 months in which to use the exemptions. It should divide the number of originally allocated average monthly exemptions by 2.06 to determine how many exemptions it could use each month and not exceed its allocation (1,650/2.06=801). The state can exempt 801 ABAWDs (23% more than the 620 that would be exempted without a time limit) but will need to end the exemption after 6 months -- in effect, giving ABAWDs who get the exemption a total of 9 months of food stamp participation when they are not working.

If Alabama implemented the provision in January, it has 9 months in which to use the exemptions. It should multiply the number of originally allocated average monthly exemptions by 12 to determine the yearly amount, and divide it by the remaining 9 months. Then it should divide that number by 2.06 to determine the number of exemptions it could use for the 9 month period from January through September ((1,650x12/9)=2,200)); (2,200/2.06=1,068). If Alabama places a 6- month limit on its exemptions as in this case, the state could exempt approximately 1,068 ABAWDs each month from January through September without exceeding its allotment.

If Alabama implements the provision in April, it has 6 months in which to use the exemptions. It should multiply the number of originally allocated average monthly exemptions by 12 to determine the yearly amount, and divide it by the remaining 6 months. Then it should divide that number by 2.06 to determine the number of exemptions it could use for the 6 month period from April through September ((1,650x12/6)=3,300)); (3,300/2.06=1,602).

If Alabama places a 6- month limit on its exemptions as in this case, the state could exempt approximately 1,602 ABAWDs each month from April through September without exceeding its allotment.

###### Table C - 3-month time limit

Again, Table C is similar to Tables A and B, except the calculations in Table C assume that the state imposes a 3-month time limit on the exemptions – in effect giving ABAWDs who get the exemption a total of six months worth of participation when they are unemployed. State agencies can use the same method if it wanted to determine how many case-months worth of exemptions it would need if it put a 3-month time limit on the exemption. However, for a 3-month time limit it would need to use 1.41 months.

If Alabama implemented the provision in October, it has 12 months in which to use the exemptions. It should divided the number of originally allocated average monthly exemptions by 1.41 to determine how many exemptions it could use each month and not exceed its allocation (1,650/1.41=1170). This limit would allow the state to give an exemption to 1,170 individuals -- 80% more persons in an average month than an exemption policy that did not impose a time limit.

If Alabama implemented the provision in January, it has 9 months in which to use the exemptions. It should multiply the number of originally allocated average monthly exemptions by 12 to determine the yearly amount, and divide that amount by the remaining 9 months. Then it should divide that number by 1.41 to determine the number of exemptions it could use for the 9 month period from January through September ((1,650x12)/9=2,200)); (2,200/1.41=1,560).

If Alabama places a 3-month limit on its exemptions and implements the exemptions in January, the state could exempt approximately 1,560 ABAWDs each month from January through September without exceeding its allotment. If Alabama implements the provision in April, it has 6 months in which to use the exemptions. It should multiply the number of originally allocated average monthly exemptions by 12 to determine the yearly amount, and divide it by the remaining 6 months. Then it should divide that number by 1.41 to determine the number of exemptions it could use for the 6 month period from April through September ((1,650x12/6)=3,300)); (3,300/1.41=2,340).

If Alabama places a 3- month limit on its exemptions as in this case, the state could exempt approximately 2,340 ABAWDs each month from April through September without exceeding its allotment.

###### Table D - Using the information for Specific Groups within the ABAWD Population

Obviously a state would use up its exemptions rather quickly if it tried to exempt every ABAWD as he/she became eligible. Even placing time limits on the duration of the exemptions would not prevent a state from exceeding its maximum number of exemptions over the course of a year. Therefore, states may want to consider further narrowing the field of those eligible for exemptions by granting exemptions to certain categories of individuals.

Table D provides information to assist states by providing estimates of the average number of ABAWDs for every 1,000 Food Stamp cases. If an area within a state has approved waivers, the state may wish to exempt ABAWDS in the remainder of the that area but is unsure how many ABAWDs there may be. However, the state may be able to estimate how many total Food Stamp cases are in the unwaived portion of the area. Table D provides the necessary information to estimate how many ABAWDs to expect per every 1,000 food stamp cases. Furthermore, Table D also provides the average number of ABAWDs for a variety of demographic and other categories per every 1,000 food stamp cases:

Age ranges within the 18-50 year old group. ABAWDs at the older end of the 18-50 age range may be more likely to face difficulties in finding the types of unskilled jobs that will enable them to qualify for food stamps than their younger counterparts.

**Persons living in metropolitan vs. non-metropolitan areas**. Some ABAWDs may also be thwarted in their efforts to find work because they live in areas that are physically remote from the places where they might find jobs.

**Educational attainment and progress toward high-school equivalency**. Research has indicated that persons who have not completed high school have a more difficult time finding even unskilled, low-wage employment than those who have completed high school.

**Non-English speakers**. Persons who do not speak English as their primary language may also have difficulty finding work regardless of their level of educational attainment.

**Persons without access to adequate transportation**. Like those who live in remote areas, persons who do not have adequate transportation will have a difficult time finding, getting to, and keeping jobs.

FNS has included information that can help states generate ABAWD caseload estimates for each of these categories of individuals in Table D. Column 1 of this table shows the average number of monthly exemptions for each state. Column 2 shows the estimated number of ABAWDs out of every 1000 people in each state’s total caseload. Columns 3-6 show the estimated number of those ABAWDs that fall within certain age groups. Columns7-9 show the estimated number of those ABAWDs who have attained various levels of education. Column 10 shows the estimated number of those ABAWDs that do not live in a metropolitan statistical area. Columns 11-12 show the estimated number of those ABAWDs that do not have vehicles.

Again, using Alabama as an example, for every 1000 Food Stamp cases, Alabama can expect to have 35 ABAWDs. If Alabama wanted to provide exemptions in a portion of the state with 20,000 cases; they can expect about 700 ABAWDs in that area (20x35). Columns 3-6 indicate that for every 35 ABAWDs, five are aged 18-20; ten are 21-31, twelve are age 31-40; and seven are aged 41-49. Column 7 indicates that 17 ABAWDs out of the group of 35 have not graduated from high school. Two out of those 35 may be pursuing a GED, and a small but indeterminate number may be non-native English speakers. Column 10 indicates that on average for those 35 ABAWDs in Alabama, there are 13 who do not live in a metropolitan statistical area (this column may not apply if the state knows for certain whether an area is urban or rural). The last two columns indicate for Alabama, 22 out of every 35 ABAWDs in urban areas, and 17 out of every 35 ABAWDs in rural areas may not have access to adequate transportation.

FNS would like to emphasize that the numbers are estimates and there are inherent weaknesses in all of them. The numbers are based on FY 1996 Food Stamp Program Quality Control (QC) data, U.S. Department of Education National Center for Education Statistics, and the Survey of Income and Program Participation (SIPP). Again, we would like to stress that while these data are the best available, they cannot perfectly capture the characteristics of the current ABAWD caseload. Much of the data predate the passage of welfare reform; it is likely that many participants who were ABAWD at that time have since lost their eligibility, and that the estimates generated with the data in Table D may overstate the current number of ABAWDs in an area. Additionally, columns 3-6 and 10 contain information that for some states, is based on fewer than 30 unweighted cases in the QC file, and therefore has a low confidence level of accuracy (the states and data for which this applies are annotated in the table footnotes). Columns 11 and 12 were generated using national level percentages, and are unable to account for the availability of public transportation in a given location.

Despite the weaknesses, states can use this information to estimate how many exemptions it would need to exempt entire categories of individuals or what percentage of a category it could exempt without exceeding the average monthly number of exemptions it has been allotted.

FNS believes that this data should help states make more informed choices about using their exemption authority. Caseload characteristics and dynamics may change in different ways in different states over time. Only by granting exemptions and measuring them can a state gain current information on how to most effectively implement its exemption policy. If the state finds that fewer exemptions than anticipated are being used, it can start giving out more. If more exemptions are being used than planned, it can scale back.