Methodology for Determining Whether an Increase in a State or Territory's Child Poverty Rate Is the Result of the TANF Program

AGENCY:

Administration for Children and Families, HHS.

ACTION:

Final rule.

SUMMARY:

This final rule establishes the methodology the Administration for Children and Families will use to determine the child poverty rate in each State and Territory. If any jurisdiction experiences an increase in its child poverty rate of five percent or more as a result of the Temporary Assistance for Needy Families (TANF) program, the State or Territory must submit and implement a corrective action plan. This requirement is a part of the TANF program, the welfare reform block grant enacted in 1996.

EFFECTIVE DATE:

This rule is effective August 22, 2000.

FOR FURTHER INFORMATION CONTACT:

Sean Hurley at (202) 401-9297 or Dennis Poe at (202) 401-4053.

Deaf and hearing-impaired individuals may call the Federal Dual Party Relay Service at 1-800-877-8339 between 8:00 a.m. and 7:00 p.m. Eastern time.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Statutory Provisions and Regulatory History

II. Provisions of the NPRM

III. Comment Overview

IV. Discussion of the Final Rule

A. Response to Comments on Cross-cutting Issues and Issues Not Tied to the Regulatory Text

1. Intent of the statute and the relationship between the TANF program and child poverty

2. Tribal TANF programs and section 413(i) of the Act

3. Recommendations to improve the national poverty measure

B. Summary of the Final Rule

C. Section-by-Section Discussion of the Regulatory Text

V. Regulatory Impact Analyses

A. Executive Order 12866

B. Regulatory Flexibility Analysis

C. Assessment of the Impact on Family Well-Being

D. Paperwork Reduction Act

E. Unfunded Mandates Reform Act of 1995

F. Congressional Review

G. Executive Order 13132

I. Statutory Provisions and Regulatory History

On September 23, 1998, the Administration for Children and Families (ACF) published a Notice of Proposed Rulemaking to implement section 413(i) of the Social Security Act (63 FR 50837). This section of the Act is a part of the welfare reform block grant program known as Temporary Assistance for Needy Families, or TANF.

The TANF program was added to the Social Security Act by the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), signed by President Clinton on August 22, 1996. The first title of this new law, “Block Grants for Temporary Assistance for Needy Families” (sections 101-116, Pub. L. 104-193) established a comprehensive welfare reform program designed to change dramatically the nation's welfare system into one that promotes work and responsibility. The new program is called Temporary Assistance for Needy Families in recognition of its focus on time-limiting assistance and moving recipients into work.

PRWORA repealed the existing welfare program known as Aid to Families with Dependent Children (AFDC), which provided cash assistance to needy families on an entitlement basis. It also repealed the related programs known as the Job Opportunities and Basic Skills Training program and Emergency Assistance.

The new TANF program went into effect on July 1, 1997, except in States that elected to submit a complete plan and implement the program at an earlier date.

This landmark welfare reform legislation dramatically affects not only needy families, but also intergovernmental relationships. It challenges Federal, State (including Territories), Tribal, and local governments to foster positive changes in the culture of the welfare system and to take more responsibility for program results and outcomes. It also challenges them to develop strong interagency collaborations and improve their partnerships with legislators, advocates, businesses, labor, community, and faith-based groups, and other parties that share their interest in helping needy families transition into the mainstream economy.

This legislation also gives States and Tribes the authority to use Federal welfare funds “in any manner that is reasonably calculated to accomplish” one or more of the four purposes of the new program. It provides them broad flexibility to set eligibility rules and decide what benefits are most appropriate, and it offers States and Tribes an opportunity to try innovative ideas so they can respond more effectively to the needs of families within their own unique environments.

One of the concerns of Congress in passing PRWORA, however, was the potential harm to children that might result from the loss of Federal entitlement to benefits or the unsuccessful efforts of their caretakers to achieve self-sufficiency within the five-year time limit for receipt of federally-funded TANF assistance. Congress was also concerned that States might take an overly-cautious approach to implementing the new law and, for example, not take advantage of the opportunities under the TANF program to use new ways to assist families to obtain and retain employment and increase economic capability.

To address these concerns, Congress added section 413(i) (42 USC 613(i)) to the Social Security Act (the Act)). Specifically:

• Section 413(i)(1) of the Act requires the Chief Executive Officer of each State (including the Territories) to submit annually to the Secretary a statement of the child poverty rate in the State. The first statement, due May 31, 1998, was required to report on the child poverty rate at the time of enactment of PRWORA, or August 22, 1996.
• Section 413(i)(2) specifies that, in subsequent years, if the child poverty rate in a State increases by five percent or more from the previous year as a result of the TANF program(s) in the State, the State shall prepare and submit a corrective action plan to the Secretary.
• Section 413(i)(3) provides that the corrective action plan shall outline the manner in which the State will reduce the child poverty rate in the State and include a description of the actions to be taken by the State under the plan.
• Section 413(i)(4) specifies that the State shall implement the corrective action plan until the State determines that the child poverty rate in the State is less than the lowest child poverty rate on the basis of which the State was required to submit the corrective action plan.
• Section 413(i)(5) requires the Secretary to establish the methodology by which a State will determine the child poverty rate and specifies three factors that the Department must take into account in developing the methodology: The number of children who receive free or reduced-price lunches; the number of food stamp households; and, to the extent available, the county-by-county estimates of children in poverty as determined by the Census Bureau.

On May 29, 1998, the Administration for Children and Families (ACF) issued a Program Instruction to States (and Territories operating a TANF program) clarifying that the State and the Territory need not submit a statement of its child poverty rate to us by May 31, 1998, as specified in the statute. We explained that we planned to send to each jurisdiction the Census Bureau estimate of the number of children in poverty and that we would be publishing an NPRM in the near future. See TANF-ACF-PI-98-4.

II. Provisions of the NPRM

Prior to development and publication of the Notice of Proposed Rulemaking (NPRM), we held two types of consultations. First, we raised issues related to this provision in the general TANF consultation meetings with representatives of State and local government; nonprofit, advocacy, and community organizations; foundations; and others. Second, we held consultations focused specifically on this provision with national organizations representing State and local elected officials; technical, statistical, and policy experts; and representatives of research, advocacy, and public interest organizations that focus on poverty and child economic well-being. These discussions were helpful to us in identifying key issues and evaluating policy options.

In the NPRM, we discussed issues raised during our consultations, including: Measurement of Child Poverty and the Census Bureau Data, Use of the County-By-County Estimates of Children in Poverty in the Methodology, Use of Food Stamp Data in the Methodology, Use of Free and Reduced-Price School Lunch Data in the Methodology, Relative Importance of Various Factors in the Proposed Methodology, and Clarification of the Term “Five Percent Increase.”

In the NPRM, our approach to establishing a methodology for determining a State child poverty rate was based on several principles: Using the most reliable and objective data on child poverty currently available (and thus avoiding a requirement that each State or Territory must develop its own child poverty rate); assuring that the child poverty rate was assessed in relation to the TANF program in the context of all appropriate circumstances in the jurisdiction; and limiting administrative burden by requiring that States and Territories provide only those data readily available and necessary to implement the statute.

We proposed a sequential methodology consisting of five major steps. Not all States or Territories would be required to participate in all steps. The proposed methodology for the Territories was similar to that for the States but included some necessary modifications.

We based our methodology on the estimates of child poverty (the child poverty rate) developed by the Census Bureau. The Census Bureau's child poverty rate is the official United States child poverty rate.

Proposed Step 1

• Annually, we would provide each State with an estimate of the number and percentage of children living at or below 100 percent of the Federal poverty threshold within the State, based on Census Bureau data. This estimate would be for the calendar year two years prior to the current calendar year, e.g., in 1998, we would provide an estimate for calendar year 1996. The Census Bureau estimates would incorporate county-level estimates of poverty.
• In 1999, and annually thereafter, we would determine for each State, at the 80-percent confidence level, the change in the percent of children in poverty for the most recent two-year period for which the data are available and provide this information to the State. In 1999, we would provide data comparing calendar years 1996 and 1997.

Proposed Step 2

• If the child poverty rate in a State did not increase by five percent or more, we would conclude that the State has met the requirements of section 413(i) of the Act, and the State would not be required to submit further information for that two-year period. (A five percent increase would mean that the most recent child poverty rate is at least five percent higher (i.e., 1.05 times higher) than the previous year's rate. A five percent increase did not mean a five percentage point increase.)
• If the child poverty rate in a State increased by five percent or more, we proposed to require that the State provide supplemental information to adjust, explain, or account for this increase. We proposed that the State, within 60 days—

a. Must provide data on the average monthly number of households with children that received food stamp benefits for each of the two most recent calendar years for which data are available;

b. Must provide data on any changes in legislation, policy, or program procedures that have had a substantial impact on the number of households with children receiving food stamp benefits during the same two-year period, including data on sub-populations affected; and

c. May provide, at State option, other information covering any pertinent time period, such as the proportion of students certified for free or reduced-price school lunches or estimates of child poverty derived from an independent source.

Alternatively, if a State chose to accept the increase in child poverty as indicated by the Census Bureau data, it could skip steps two and three and move directly to step four—the assessment of the impact of the TANF program on the increase in child poverty.

Proposed Step 3

• We would review the food stamp and other data provided by the State. If we determined that these data indicated a subsequent improvement, commensurate with the poverty increase in the Census data, it would not be necessary for the State to proceed to Step 4 because the more recent data would indicate that the child poverty rate in the State was improving.

Proposed Step 4

• If we determined that the food stamp and other data provided by the State did not indicate a subsequent commensurate decrease in child poverty, we proposed to notify the State that it must, within 60 days, provide an assessment (and the information and evidence on which the assessment was based) of the impact of the TANF program in the State on the child poverty rate. We proposed to give the States and Territories broad latitude in the information they could provide.

Proposed Step 5

• We would review the information provided by the State, along with other information available such as the State's TANF plan and eligibility criteria, data on other supportive services and assistance programs, and information on the State's economic circumstances. If we determined that the increase in the child poverty rate was the result of the State's TANF program, we would notify the State that it would be required to submit a corrective action plan within 90 days.

Proposed Methodology for the Territories

• To the extent that data are available and the procedures applicable, we proposed that the Territories would be subject to the same general methodology as described for the States. Because the Census Bureau does not estimate a child poverty rate for the Territories, we proposed that ACF would compute an estimated child poverty rate for each Territory, based on information submitted by the Territory.
• Subsequent procedural steps would be the same as for States, i.e., as applicable, we would review supplemental data to determine whether the child poverty rate increased by five percent or more; review the Territory's assessment of whether the increase in the child poverty rate was a result of the TANF program; and require the development of a corrective action plan, as necessary.

Based on this proposed methodology, we anticipated that a small number of States and Territories would need to respond to the requirements of each step and an even smaller number would be required to submit a corrective action plan.

III. Comment Overview

We received 14 comment letters on the NPRM from seven State TANF agencies, four national organizations, two State and local policy and advocacy organizations, and one United States Senator. We reviewed and seriously considered all comments. We particularly appreciated the fact that several commenters went beyond reacting to the proposed regulatory text to include a helpful discussion of the issues raised in the preamble and additional supportive and analytic information.

In general, most commenters had mixed views on our proposed approach. They commended our external consultation process prior to the development of the rule and our “reader-friendly” regulations, given the highly complex and technical nature of the subject. Several commenters agreed with specific policy provisions, e.g., our use of the Census Bureau data, our recognition of the limited usefulness of the school nutrition program data, and the flexibility we proposed to allow States regarding what information the State could include in its assessment of the impact of the TANF program on child poverty or in the corrective action plan.

At the same time, we also received some objections to our proposed approach and recommendations for changes. The strongest objections were directed at our proposal to allow a State whose child poverty rate had increased by five percent or more to provide food stamp participation data in order to adjust for deficiencies in the Census Bureau data. Our rationale in the NPRM was that food stamp participation data (which historically had tracked the poverty rate) could be used to show evidence of more recent trends that would explain or “rebut” the increase in child poverty. Commenters pointed out that the food stamp participation rate, indeed, had tracked poverty in the past but that recent evidence indicated that it no longer did so. They urged the deletion of this provision. Others objected to this provision on the grounds of administrative and reporting burden.

Two commenters objected to what they believed were implicit assumptions in the statute, i.e., that child poverty is the result of the TANF program or that the TANF program could affect child poverty in any meaningful way. Others objected to the additional administrative burden of specific provisions and questioned several technical provisions, e.g., our use of the 80-percent confidence interval in determining the child poverty rate.

Some commenters called to our attention that we had not addressed the role of the Tribal TANF programs in implementing this section of the Act. Some recommended that we clarify that States may exclude the Tribal TANF population in the calculation of the State's child poverty rate.

In addition, one advocacy organization urged us to focus not only on a five percent increase in the child poverty rate but also to address the “poverty gap,” i.e., the depth of poverty for those children below the poverty level. Finally, two national organizations recommended a number of steps the Department might take to help improve the national child poverty measure and, thus, better implement the overall intent of the statute.

We have organized our response to the comments, first, to address the issues that are cross-cutting and are not tied to regulatory text and, second, to address other issues in the section-by-section discussion of the regulatory text.

IV. Discussion of the Final Rule

A. Response to Comments on Cross-cutting Issues and Issues Not Tied to the Regulatory Text

1. The Intent of the Statute and the Relationship Between the TANF Program and Child Poverty

Comments: A number of commenters expressed differing views on the purpose of section 413(i) of the Act and the NPRM. One commenter assumed that the purpose of the statute and the regulation was to decrease child poverty nationally. Other commenters believed that the intent of the law was to monitor child well-being and track changes in the child poverty rate related to PRWORA. Some commenters objected to what they believed were implicit assumptions in section 413(i), i.e., that the child poverty rate was the result of the TANF program (a “cause and effect relationship” was assumed to exist), or that the TANF program could affect child poverty in any meaningful way. One commenter found the statute and the NPRM “grossly flawed” based on this implicit assumption.

One commenter stated that, as was true in the Aid to Families With Dependent Children (AFDC) program, the TANF program is not explicitly designed to elevate families above the poverty level. Rather, its purpose, they believed, is to provide a set of financial and service supports, coupled with an assumption of personal responsibility, that will provide the opportunity for a family to become self-sufficient. Except in circumstances where the State's TANF payments exceed the poverty rate, this commenter alleged that all children receiving TANF will already have incomes beneath the poverty level.

Therefore, because all the affected persons are already counted as living beneath the poverty level, no change to the operation of the TANF program, whether it be reducing TANF payments, failing to move families to employment, or terminating families' eligibility for TANF, would increase the poverty rate.

They concluded that only positive changes made by the TANF program, such as successful employment programs which move recipients to relatively high paying jobs, could affect the child poverty rate. Significant changes in the poverty rate, they believed, are necessarily the result of factors extrinsic to TANF, such as economic and demographic shifts. Thus, it appeared to this commenter that neither the statute nor the regulations could be implemented in any meaningful way.

Response: We disagree with the observation that “all children receiving TANF will already have incomes below the poverty line.” Based on AFDC data, we know that, typically, a family may have income below the poverty line in a specific month (or months), but family income would not necessarily fall below the poverty line on an annual basis.

However, we agree that the intent of the statute reflects Congressional concern about PRWORA's effects on the well-being of children, including children who no longer receive TANF benefits. Section 413(i), as well as other provisions of the law, were added to provide a careful look at what is happening to children following enactment of this legislation. Clearly, certain TANF program and policy decisions could contribute to an increase in poverty. Finally, and more importantly, we note that there are positive actions that State and Tribal TANF agencies can take to help improve the poverty status and well-being of families.

The child poverty rate in the United States, developed by the Census Bureau, is a frequently used indicator of child well-being. (The “child poverty rate” means the percentage of all children in a State that live in families with annual incomes below 100 percent of the Census Bureau's poverty threshold.)

The national child poverty rate has declined since 1992 from 22.3 percent to 18.9 percent, the largest five-year drop in nearly 30 years. Still, currently, 13.5 million children live below the poverty line. At the same time, a recent Census Bureau report found that, among Americans living below the poverty line, a greater share held jobs than at any point in the last 20 years. The Census Bureau found that, in 1998, 12.5 percent of poor adults worked full-time (a 22 percent increase over 1997), and another 41 percent worked part-time.

In the context of the TANF program, employment is central to assisting families to escape poverty. States have made huge progress in moving families to work; large increases in employment are evident from every information source. However, for many families, work by itself will not guarantee an escape from poverty unless other critical supports are in place. Thus, the challenges are to continue the movement of families into work, build supports that can sustain parents in work, and help them move to more enduring and higher paying jobs, so that families who work will not be poor.

The TANF program, as opposed to the AFDC program, allows States to provide a broad array of supports for working families and to provide them independently of the basic cash welfare system. Unlike AFDC, TANF can be an effective vehicle for reducing poverty, supporting families, and making work pay.

A number of innovative States are using child poverty as a measure of their efforts to assist families, and some States are already using the resources and flexibility under TANF to address this issue. Some activities that specifically address poverty include:

• Under TANF, utilizing well-known strategies to supplement work, such as more generous earning disregards, earnings supplements, and wage subsidies;
• Improving child support, such as increasing the amount of support collected from non-custodial parents that is passed through to children; and
• Enacting State refundable tax credits.

Recent research findings from studies in Minnesota and Oregon support the use of these specific strategies in reducing poverty.

In addition to these activities directly related to reducing poverty, States are undertaking a number of supportive activities which indirectly help make work pay, including:

• Taking critical steps to ensure that eligible families, including those that do not receive TANF, do receive food stamps and Earned Income Tax Credit payments for which they are eligible;
• Increasing the stability of work through investments in the wages parents earn or the hours they work, such as employer partnerships that focus on the first job, on job advancement after the first job, or on combinations of work and training; mentoring and case management strategies; strategies that combine work, education, and training; and supported work for families with barriers to private sector employment;
• Helping families during periods between jobs, such as quick re-employment services; and
• Providing employment assistance for other families, such as a child-only family where a caretaker relative is not receiving assistance.

We are continuing to monitor what is happening to children and families as a result of the enactment of the TANF program. In addition to section 413(i), we are looking at State performance and accomplishments through the High Performance Bonus and the Out-of-Wedlock Childbearing Bonus. We are also sponsoring a variety of research studies and evaluations to assess the impact of welfare reform, e.g., we are measuring the effects of different approaches to welfare reform on child well-being, and numerous studies are tracking families leaving TANF.

2. Tribal TANF Programs and Section 413(i) of the Act

Comments: As several commenters pointed out, we did not address the issue of child poverty in areas covered by Tribal TANF programs in the NPRM. They asked for clarification in the final rule on how Tribes operating TANF programs will be considered in the poverty rate calculation and recommended that we allow States to exclude the Tribal TANF population in the calculation of a State's child poverty rate. These commenters also indicated that it was unfair to hold the State TANF program accountable for the Statewide child poverty rate when the State has no authority over or responsibility for the conduct of the TANF program in areas of the State covered by a Tribal TANF program(s).

Response: Section 413(i) of the Act specifies the responsibilities of the Chief Executive Officer of the State in relation to increases in the child poverty rate and the TANF program(s) in the State. Section 413(i)(2) also provides that an assessment of the increase in the State's child poverty rate shall be made in relation to “the amendments made by section 103 of PRWORA.” Because section 103 of PRWORA authorizes both State and Tribal TANF programs, the Chief Executive Officer must address increases in the State's child poverty rate in relation to both State and Tribal TANF programs in the State.

We do not accept the recommendation that the States may exclude the Tribal TANF population in a calculation of a “State” child poverty rate. We could not implement this recommendation because the statute clearly specifies that both State and Tribal TANF programs must be considered. In addition, the Census Bureau does not determine a separate child poverty rate for Tribal lands or reservation areas.

In response to comments, however, we have amended three sections of the final rule. Specifically, we have:

• Amended § 284.15(b) to provide that the State should obtain information from and work with any Tribe(s) (and Tribal consortia) operating a Tribal TANF program in the State in preparing and submitting the assessment of the impact of TANF programs on the increase in child poverty and the corrective action plan;
• Added, in § 284.30(b), examples of Tribal TANF information that might appropriately be included in the assessment;
• Extended, in § 284.21, the period of time the State has submit the assessment, from 60 days to 90 days, in part to allow further opportunity for State and Tribal coordination; and
• Specified in § 284.45 that any actions to reduce child poverty to be taken by the Tribe(s) must be included in the corrective action plan.

In the context of State and Tribal cooperation, we note that there are 330 American Indian entities in the contiguous 48 States and 13 Alaska entities eligible to administer a Tribal TANF program. Currently, there are 21 approved Tribal TANF plans in operation. (One additional Tribal TANF plan is approved but not yet in operation.) Nineteen of these programs involve individual Tribes and three are operated by inter-tribal consortia. (One consortium in California is composed of 19 Tribes; another consortium is composed of 37 Alaska Native villages (Tribes); and the third consortium is made up of 20 Alaska Native villages (Tribes).) Additional Tribes are exploring the option of operating a TANF program.

Both State and Tribal TANF programs serve Indian families. Based on the most recent data available from 1999, Indian tribes expect to serve approximately 3,800 families in FY 2000. In FY 1998, approximately 46,702 American Indian families were served by State programs. In several States, American Indians constitute a large percentage of the State TANF caseload, i.e., 73 percent of South Dakota's TANF caseload, 54 percent in North Dakota, about 41 percent in Alaska, and over 46 percent in Montana.

Tribes that operate a TANF program have the flexibility to design programs and services; define eligibility criteria; establish benefits; and design strategies for achieving program goals, including helping recipients become self-sufficient.

Welfare reform also provided Tribes and States with new opportunities for communication, coordination, and collaboration to help achieve program goals. One of the most important ways States have been working with Tribes to address the issue of poverty is by making State supplemental contributions to Tribal TANF programs (as a maintenance-of-effort (MOE) expenditure) to expand job-related activities and strengthen Tribal programs for families and children. In addition, a number of States and Tribes are also entering into a range of cooperative efforts, including:

• Sharing equipment and resources, such as computers;
• Co-locating service centers and sharing office space;
• Conducting joint staff training;
• Coordinating information, reporting requirements, and reporting systems;
• Establishing consolidated intake and eligibility determinations, particularly for the food stamp, Medicaid, and State Children's Health Insurance Programs; and
• Cooperating in the provision of direct services (e.g., job skills training) and supportive services, e.g., transportation.

The preparation and submittal of an assessment and a corrective action plan are additional opportunities for State and Tribal coordination, both in meeting the requirements of section 413 of the Act and meeting the needs of Indian families. We believe the additional time provided to submit the assessment will help support such coordination.

3. Recommendations To Improve the National Poverty Measures

Comment: Two national advocacy organizations recommended that the U.S. Department of Health and Human Services (DHHS) should take full advantage of the opportunities afforded by PRWORA and section 413(i) of the Act to actively explore activities that would expand and improve the Census Bureau's existing measure of child poverty. They believed that the most important purpose of section 413(i) of the Act is to require a careful look at what is happening to the well-being of children following enactment and implementation of TANF. The measurement of child poverty, therefore, provides a useful means of evaluating changes in child well-being at the State level. Improved measures will support this effort.

Generally, these organizations recommended that the Census Bureau's child poverty estimates include data from both current and new sources and that currently available data from other sources should be used (in conjunction with the official poverty measure) to focus increased public attention on child poverty. Specifically, they recommended that DHHS:

• Support research now underway to implement the recommendations of the National Academy of Sciences panel on poverty measurement;
• Support funding needed to field the Census Bureau's American Community Survey which is designed to provide more timely State-level data than the decennial census and more accurate State-level data than the Current Population Survey;
• Explore ways to use currently available data to improve the existing poverty measure and to add new data to the measure as they become available, e.g., modify the Census Bureau's Small Area Income and Poverty Estimates by adding data on food stamps, housing benefits, earned income tax credits, and work expenses;
• Encourage States to better assess the well-being of their children and make the data more generally useful by participating in the increased costs of expanding the Current Population Survey (CPS) sample size at the State level;
• Encourage and fund efforts by States to develop administrative databases for measuring child well-being within their own jurisdictions;
• Use data already collected by the Census Bureau to show the impact of specific programs such as TANF by comparing child poverty before and after receipt of means-tested government transfers; and
• Publish measures of the poverty gap among children to provide an indication of the depth of poverty, at least at the national level.

Response: While we generally agree that these recommendations would help to improve the Census Bureau's measurement of child poverty and understand the circumstances of children in poverty, the Department already participates in a number of inter-agency and Departmental efforts that address these recommendations:

• In May 1995, the Panel of Poverty and Family Assistance appointed by the National Research Council published a report in which it proposed a new approach for developing an official measure of poverty in the U.S. Since that time, personnel from the Census Bureau, the Bureau of Labor Statistics, and other Federal agencies (including DHHS) have been engaged in research to explore possible implementation of the Poverty Panel's recommendations.
• In 1997, the Office of Management and Budget convened a Federal Interagency Technical Working Group to Improve the Measurement of Income and Poverty that includes representatives from DHHS. In July 1999, the Census Bureau issued a report on experimental poverty measures, reflecting the results of ongoing research. However, none of these experimental measures has been selected to replace the current official definition because a number of issues remain unresolved. The review of alternative poverty measures is expected to carry on for several years. We will continue to work with this group and other interested public and private organizations to develop improved measures of child poverty.
• Also at an inter-agency level, the National Science and Technology Council's “Children's Initiative Subcommittee” continues to explore the most effective use of Federal resources for research focused on child poverty as well as other issues related to the well-being of America's children.

In addition, the Department is:

• Transferring funds to the Census Bureau to allow for the expansion of State and local estimates of poverty to include children ages 0 to 4;
• Financing a research effort to advance State Child Indicators Initiatives. The aims of this program are to: (1) Promote State efforts to develop and monitor indicators of the health and well-being of children as welfare reform and other policy changes occur; and (2) help to institutionalize the use of indicator data in State and local policy formulation; and
• Funding two national poverty research centers: The Institute for Research on Poverty located at the University of Wisconsin at Madison and The Joint Center for Poverty Research, co-located at the Northwestern University and the University of Chicago. These national nonprofit, non-partisan centers focus their research on the causes and consequences of poverty and inequality and on interventions to reduce poverty and dependence and help focus on and contribute to the knowledge about this important issue.

In response to the recommendation for publication of information on poverty, particularly the poverty gap, these data are available from the Census Bureau's Internet site (www.census.gov) along with information on how to obtain more detailed data files. ACF also publishes poverty gap information in its annual TANF Report to Congress.

B. Summary of the Final Rule

Our principles, first established in the development of the NPRM, remain the same for the final rule: Use the most recent, reliable, and objective data available; assess the impact of the TANF program(s) in the State on any increase in the child poverty rate of five percent or more in the context of all appropriate information; and require minimal administrative burden on States and Territories in carrying out these requirements.

In the final rule, we retained some policies as proposed in the NPRM and made several changes and modifications, based on our consideration of public comments. We address these policies and changes in the section-by-section discussion. Briefly, however, we:

• Continue to base the State child poverty rate on the current Census Bureau estimates, but, if better Census Bureau data become available, we will use these data;
• Will provide to States, in 2000, the Census Bureau's estimate of the number and percent of children in poverty in each State for calendar year 1997 and the change in the percentage of children in poverty between 1996 and 1997, at the 90-percent confidence level rather than the 80-percent confidence level proposed in the NPRM. (We provided the calendar-year 1996 Census Bureau estimates of children in poverty in each State to the Chief Executive Officer of the 50 States and the District of Columbia on December 21, 1999. We also sent a copy of this information to the director of each State's human services agency.);
• Allow a State to submit an estimate of the State's child poverty rate, derived from an independent source; we will accept the State's estimate if it is more reliable than the Census Bureau data;
• Eliminated the step in the NPRM that used food stamp participation rate data to “rebut” an increase in the child poverty rate;
• Continue to allow a State wide latitude in how it conducts its assessment of the impact of the TANF program(s) in the State on an increase in the State's child poverty rate;
• Allow a State 90 days to submit the assessment, an additional 30 days beyond the 60-day period proposed in the NPRM;
• Continue to allow States to develop the content and determine the duration of the corrective action plan in accordance with the law;
• Clarify that a State should obtain information from and work with the Indian tribes (or Tribal consortia) operating a TANF program in the State in preparing and submitting the assessment and the corrective action plan; and
• Postpone the development of a child poverty rate for the Territories (i.e., Guam, Puerto Rico, and the Virgin Islands) until reliable data are available. (At the present time, American Samoa has not applied to implement the TANF program.)

C. Section-by-Section Discussion of the Regulatory Text

Section 284.10—What Does This Part Cover?

In the NPRM, this section provided a summary of the scope and content of part 284. We received no comments on this section.

We made two editorial changes in the final rule for clarity. First, we added the word “Territories” to make explicit that this part applies to States and Territories. Second, we substituted the phrase, “as a result of the TANF program(s) in the State or Territory” for the phrase “as a result of TANF.” This latter change emphasizes that an increase in the State's child poverty rate will be assessed in relation to all TANF programs(s) in the State, i.e., both State and Tribal TANF programs.

Section 284.11—What Definitions Apply to this Part?

This section of the NPRM proposed the definition of terms we used in part 284. We received one comment on this section indirectly related to our definition of “children in poverty.” We had defined this term to mean “estimates resulting from the Census Bureau methodology of the percent of children in a State that live in families with incomes below 100 percent of the federal poverty level.” These estimates, developed by the Census Bureau, constitute the official U.S. child poverty rate.

In our external consultations prior to the development of the NPRM, we noted that several agencies and organizations recommended that, in addition to statutory requirements, we also focus on the percent of children in families with income levels at or below 50 percent of poverty. We considered regulating beyond the provisions of the statute, but found that the current Census Bureau methodology would require significant revision and would be much less effective in estimating poverty at levels lower than 100 percent. However, we invited public comment about the advisability and desirability of pursuing such an approach.

Comment: The advocacy organization which responded to this issue pointed out that measuring a five percent increase in the child poverty rate, as required by law and as proposed in the definition of children in poverty in the NPRM, will not present a complete or accurate picture of the effects of TANF on poor children. They were concerned that changes brought about by State TANF policies could negatively impact the lives of children whose families were already below 100 percent of the poverty level before the TANF program began. They provided two examples to illustrate this point:

• In States whose pre-TANF cash assistance standards were below the federal poverty level, children in TANF families were already included in the State's poverty rate. Reductions in the amount of assistance caused by TANF-related changes, or even failure of the assistance standard to keep pace with inflation, would worsen the family's poverty but would not be reflected as a change in the poverty rate and, thus, not measure the impact of the TANF program on children and families.
• The flexibility offered to States under the TANF program means that a State could make major policy changes that might negatively affect families. For example, a State might count the benefits from other programs (such as Supplemental Security Income (SSI)) as available income. Such a policy could make a family ineligible for TANF. In this case, a family could go from receiving low assistance to no assistance and still not be identified as having been affected by TANF-related changes.

This organization was also concerned about the inequities of the proposed five percent increase and our definition. For example, in States where the poverty rate is high, a five percent increase means more children have fallen into poverty before corrective action is taken than in States whose initial poverty rates are low.

They strongly recommended that we develop measures that would not only identify the number of children in families below the poverty level, but would also reveal the depth of their poverty, i.e., the size of the “gap” between family income and 100 percent of the official poverty level. They also recommended that the final rule focus on any State where the child poverty rate is high, regardless of whether that rate increased by five percent or more.

Response: We carefully considered these comments in developing the final rule. We agree with this commenter that child poverty is a serious issue and that poverty at the deepest levels is an even more serious issue. We are committed to, and concerned with, the well-being of children and families, and undertake a wide range of activities to improve the lives of children and families, as do a number of other public and private agencies and organizations. However, given the specific requirements of the statute and our lack of regulatory authority to impose requirements not specified in the law, we did not accept this recommendation. For the purpose of public information, as noted earlier, we publish the poverty gap information in our Annual TANF Report to Congress. We also post the Census Bureau's State child poverty rate data on our Internet site (www.acf.dhhs.gov).

Other Changes Made in § 284.11 of the Final Rule

We made several changes in this section to provide further clarity and explanation of terms. We:

• Revised the definition of “Census Bureau methodology” by: (1) Adding an explanation to clarify that the term may include a range of mechanisms to estimate poverty, including estimates based on the Current Population Survey; the Small Area Income and Poverty Estimates; annual demographic programs, including the American Community Survey; or any other methods used by the Census Bureau; (2) adding a definition of the term “children in poverty” as used in the definition of “Census Bureau methodology” and deleting the definition of “children in poverty” as a separate definition; and (3) deleting the sentence explaining how we proposed to compute the child poverty rate for the Territories;
• Revised the definition of “Child poverty rate” to incorporate language from the NPRM's definition of “children in poverty” and to comport with the current Census Bureau description;
• Added a definition of Tribal TANF program; and
• Added definitions for, or explanations of, acronyms used in the final rule, i.e., the Small Area Income and Poverty Estimates (SAIPE), Separate State Program-Maintenance of Effort (SSP-MOE), and Maintenance of Effort (MOE).

Section 284.15—Who Must Submit Information to ACF To Carry Out the Requirements of this Part?

As specified in section 413(i) of the Act, we proposed in the NPRM that the Chief Executive Officer of the State or Territory, or his or her designee, is responsible for carrying out the requirements of this part.

We received no comments on this section of the NPRM but have made one addition in this section of the final rule. In new paragraph (b), we specify that the State should obtain information from and work with any Indian tribe (and Tribal consortia) operating a TANF program in the State in preparing and submitting the assessment (as specified in § 284.30) and the corrective action plan (as specified in § 284.45). As discussed above under the topic “Tribal TANF programs and Section 413(i) of the Act,” this change clarifies that the statute requires the State to consider both State and Tribal TANF programs in carrying out the requirements of this part. If the requested Tribal TANF information is not made available to the State, any submission to us should clearly indicate that fact.

Section 284.20—What Information Will We Use To Determine the Child Poverty Rate in the State?

(§ 284.20 of the NPRM—What information will we provide to each State to estimate the number of children in poverty?)

In the NPRM, we proposed, in paragraph (a), that we would send to the States each year the Census Bureau's estimate of the number of children in poverty. The first estimate in 1998 would be for calendar year 1996. In paragraph (b) of the NPRM, we proposed that, beginning in 1999, we would compute the change in the percentage of children in poverty, at the 80-percent confidence level, and provide this information to the State. We proposed that the first percentage change would be sent to States in 1999 and would cover the change between calendar years 1996 and 1997. We also proposed that the annual Census Bureau estimates would be based on the Current Population Survey (CPS) data and would incorporate data from the Small Area Income and Poverty Estimates (SAIPE), e.g., State and county level data.

We have continued this general approach in § 284.20 of the final rule, with two modifications. The first modification, in response to comments, is a change in the level of statistical confidence we will use to determine the percentage change in a State's child poverty rate, i.e., from 80 percent to 90 percent. We have also clarified the use of a statistical test to ascertain a change in a State's child poverty rate. The second modification expands on a provision in the NPRM to allow a State to submit child poverty data derived from an independent source.

Briefly, we will continue to send annual estimates of the number and percentage of children in poverty to each State, based on data from the Census Bureau. Paragraph (b) of the final rule specifies that, in 2000, we will determine the first percentage change in the State's child poverty rate, between calendar years 1996 and 1997, at the 90-percent confidence level. Paragraph (c) allows a State to submit child poverty data derived from an independent source as an alternative to the Census Bureau data and specifies the conditions for submitting these data. Paragraph (d) specifies that if we determine that the State's independent child poverty data are more reliable than the Census Bureau data, we will accept them. These changes are discussed more fully below.

We also received a number of technical comments which we have organized into and will respond to under four subject areas: Use of the Census Bureau Estimates of Poverty; Use of the 80-percent Confidence Level; Interpretation of the Term “five percent increase;” and Dissemination of the Census Bureau Data.

Comment: One national organization recommended that, because the Census Bureau data is considered moderately reliable and we proposed an 80-percent confidence level, the final rule should allow States to challenge the Census Bureau estimates by providing alternative statistical evidence.

Response: In the NPRM, we proposed in § 284.25(c) to allow a State to submit an independent estimate of child poverty as part of the “rebuttal” process, along with food stamp and school nutrition data. However, we agree with the thrust of this commenter's recommendation and have added new paragraphs (c) and (d) to § 284.20 of the final rule to allow States to provide child poverty data derived from an independent source.

New paragraph (c) specifies that if the State submits an independent estimate of child poverty, it must do so within 45 days of the date the State receives the Census Bureau estimates from us; include the child poverty rate for each of the two years covered by the Census Bureau estimates; include a computation of the change in the child poverty rate over the two-year period at the 90-percent confidence level; and provide a description of the methodology used by the independent source to develop its child poverty estimate.

New paragraph (d) provides that we will accept the State's independent estimate of the child poverty rate if the data are more reliable than the Census Bureau data. Otherwise, we will determine the State's child poverty rate based on the Census Bureau estimates.

In the NPRM, we recognized that a growing number of States and other organizations are conducting studies of child poverty. One of our aims in implementing section 413(i) of the Act is to use the best child poverty data available. We believe that those States currently conducting or funding such studies of child poverty will be in the best position to provide an independent estimate, if they so choose.

We have specified that the State may provide an alternative estimate of the State's child poverty rate, but only if the estimate is derived from an independent source. (An “independent source” is a source of data or information not under the direct supervision or control of the State TANF agency such as a university, research or advocacy organization, or an independent evaluation or analysis office associated with a State executive branch agency or State legislature.)

We have specified that the independent estimate must be provided within 45 days of the State's receipt of the Census Bureau estimates because we want to utilize independent estimates that have already been conducted for the applicable years. Our intent is to not delay the process of review and assessment of poverty estimates in relation to the TANF program(s) in the State.

We will need certain information from the State in order to evaluate the reliability of the State's independent estimate. Accordingly, in paragraph (c)(2)(iv), we specify that the State must describe the methodology used to develop the independent estimate, the source of the data, the data collection methodology, any known problems associated with making estimates of this type, the estimate of the standard error, and the power of the sample to detect a five percent change in the child poverty rate. The State must also use the official definition of poverty used by the Census Bureau.

We believe that the State's data must be “more reliable” than the Census Bureau data. Otherwise, we will use the Census Bureau data in implementing this part. For the purposes of paragraph (d), the term “more reliable” means data that are based on and meet accepted statistical methods and standards, e.g., the data are derived from a representative sample of households, determined at precision levels higher than the Census Bureau data, and based on income and other variables comparable to the Census Bureau methodology.

A. Use of the Census Bureau Estimates of Poverty

Comments: A national organization supported our use of the Census Bureau estimates on the grounds that “the Census Bureau estimates, including the SAIPE data, are the best current available measures of the percentage of children living at or below the Federal poverty threshold.” A State TANF agency expressed appreciation for our proposal to send the child poverty data to the States, thus reducing State reporting burden.

However, another State TANF agency recommended that we use only the Current Population Survey (CPS) data because the CPS sample sizes are large enough to reduce the risk of error, higher confidence levels are possible, and the lag time would be reduced from two years to one.

Response: We reviewed our decision to use the Census Bureau estimates of poverty, and we believe, at the present time, they are the best national data available. This decision is reflected in paragraph (b) of this section of the final rule. If more reliable Census Bureau data sources become available in the future, we plan to use them.

In response to the recommendation that we use CPS data as the basis of our national estimates of poverty, we agree with the commenter that the lag time would be reduced. However, we believe that the SAIPE data are not only required by section 413(i)(5) (i.e., use of “county-by-county estimates of children in poverty as determined by the Census Bureau”), but that the SAIPE data, when used with the Census Bureau State estimates, provide greater accuracy and less variation than are present in the CPS data.

In addition, we are not relying on point estimates, but are using statistical tests that address variation. Finally, we believe, and the Census Bureau confirms, that the CPS estimates are large enough to provide reliable direct estimates for only 10 States and a few large counties. The CPS data might serve the large population States but would not serve as a reliable national data source for all States.

Comment: One commenter noted the increasing disconnect between the use of food stamps and the poverty rate and was concerned about how this would affect the SAIPE data.

Response: The Census Bureau is satisfied that current estimates are reasonable and appropriate. They will be closely monitoring the relationship between food stamp program participation and poverty and will consider changes in their modeling process, as needed. In addition to the Census Bureau's expert review, a national panel of independent experts, established by Congress under the auspices of the National Academy of Sciences' Committee on National Statistics, has been formed to also review and determine the reliability of these estimates. The results of this independent review are available from the Academy on its website.

Comment: A State TANF agency requested information concerning how the Census Bureau determines the child poverty rate, the formula, weighting of variables, and the definition of child poverty used. They believed this information would be useful in order to monitor and modify their TANF program to avoid negative impacts.

Response: We refer this commenter to the Census Bureau website. On its Internet site, the Census Bureau provides a wide range of information on families and children in poverty. It also provides an explanation of its methods, discusses the limitations of and problems encountered in its use of current methods, and describes steps being taken to improve future data collection.

B. Use of the 80-Percent Confidence Level

Background

The measurement of child poverty involves a process which employs samples taken from the general population to generate estimates. In this case, we use sample estimates produced by the Census Bureau to determine if an increase in the child poverty rate has occurred over a two-year period. This process of using samples results in some statistical uncertainty in each year's estimate of the child poverty rate. It is because of this statistical uncertainty that we cannot simply look at the difference in the observed poverty rate from one year to the next and determine that an increase of at least five percent occurred or did not occur. To overcome this statistical uncertainty, statisticians employ tests that incorporate a measure of error to better estimate a characteristic of a population.

In the NPRM, we proposed to use a statistical test at the 80-percent confidence level. We proposed the 80-percent confidence level for several reasons. First, we were attempting to strike a balance between falsely identifying a State as having a five percent increase in its child poverty rate when it did not and missing a five percent increase that truly did occur.

Second, we proposed that if a five percent or greater increase in the poverty rate had occurred, based on the Census Bureau's data and the statistical test, we asked that the State submit more recent Food Stamp administrative data to determine if there was a commensurate increase in the number of households with children participating in the Food Stamp Program. Finally, we reasoned that the direct consequence to a State of the identification of a five percent increase when in fact such an increase did not occur would be the preparation of an assessment of the impact of the TANF program on the increase and, possibly, a corrective action plan.

In the NPRM, we proposed a one-tailed statistical test because we were interested in determining if an increase in the child poverty rate occurred from one year to the next. There are three possible outcomes from comparing the child poverty rate from one year to the next: (a) There was an increase; (b) there was no change; and (c) there was a decrease. However, only an increase is relevant for purposes of this section of the Act. We, therefore, proposed using a one-tailed statistical test which only examines increases or decreases, but not both.

Comments: One national advocacy organization supported our use of the 80-percent confidence level, concurring with our statement in the NPRM that it would help protect children by decreasing the possibility that we would miss a significant change in a State's child poverty rate. They also supported the 80-percent confidence level as a useful tool in carrying out what they believed was the primary intent of section 413(i), i.e., the protection of children.

Four State TANF agencies, however, disagreed with our proposed use of the 80-percent confidence level. Their general concern was that the lower confidence level would lead to a large number of States incorrectly identified as having experienced a five percent increase in child poverty. One State agency noted that, using an 80 percent confidence level, we could expect that, purely by chance, 10 (or 20 percent) of the States would show an increase in poverty when in fact they do not. At the 90-percent level, five States could be mistakenly identified as having an increase in poverty.

Another State agency observed that any State whose child poverty rate increased by five percent or more could maintain that the larger confidence level obscured any real fluctuation. In other words, an alleged increase could be the result of random fluctuation.

Still another State agency commented that the use of a one-tailed confidence interval test does not appear to take into account the error in both the previous year and the current year estimate.

Recommendations by these State agencies included:

• Increase the confidence level to 90 or 95 percent because these levels are used in much policy research, including the national welfare reform demonstrations sponsored by DHHS.
• Use, and clearly explain in the final rule, a statistical test that appropriately provides for the error in the previous and the current year estimates, e.g., a difference of means or proportions test or a confidence interval around the difference in proportions.
• Make explicit in the final rule the possibility, magnitude, and the benefits of an incorrect identification of an increase in the child poverty rate.
• Explicitly state in the final rule the tradeoffs in choosing a particular confidence level, including the use of a “payoff matrix, computable by a statistician, using the standard errors from the model that will be used.”

Response: After consideration of all comments on both the policy and technical issues, we have concluded that the intent of the statute and this section of the law will be served if we increased the level of statistical confidence from 80 to 90 percent while maintaining the use of a one-tailed Z-test for the difference of proportions. We believe that this choice will serve the needs of children while balancing the burden on the States.

First, we concur with the national advocacy organization that the use of the 80-percent confidence level is a useful way to assess the change in the child poverty rate for the purpose of the Act. At the same time, the State TANF agencies are correct in noting that, at the lower confidence level (80 percent as opposed to 90 or 95 percent), we can expect a larger number of findings of change that may be due to statistical variability and not true increases. Increasing the confidence level to 90 percent would reduce the likelihood that a State would be identified as having an increase in its child poverty rate when in fact the rate had not increased. We are aware, however, that this change leads to the possibility that we may miss a State where an increase actually did occur. However, on balance, we believe this change will provide more reliable data while protecting the well-being of the nation's children.

Second, we want to make explicit the statistical test we will employ to determine if a five percent or greater increase in a State's child poverty rates occurred. We use a Z-test for difference between proportions. This test uses the information for each year under consideration, including the point estimate of the child poverty rate for each year as well as the variance of the point estimate for each year. In addition, since the samples for the two years are not completely independent, the statistical test utilizes the variance of the difference between the point estimates for the two years.

Given that we have eliminated the step that proposed to rely on Food Stamp participation rate data to “rebut” an increase in the child poverty rate, a State found to have an increase of five percent in its child poverty rate would either accept the finding and provide an explanation of the role of TANF in the increase, or provide child poverty data from an independent source as an alternate to the Census Bureau's estimates.

C. Interpretation of the Term “Five Percent Increase”

In the NPRM, we proposed that the term “five percent increase,” as specified in the statute, did not mean a five percentage point increase in child poverty. Rather, it meant that the most recent child poverty rate is at least five percent higher than the previous year's rate.

Comment: One State TANF agency, in commenting on the statistical concerns they had with our approach and the consequent burden on States of this approach, noted that there were “high risks” involved in trying to detect a small percentage change in the face of large errors in the estimation procedures. They asserted that Congress would not have wanted to impose a significant corrective action burden on States based on erroneous data. They believed Congress must have meant an increase of five percentage points.

Response: Based on our analysis, we believe that there is enough reliability in the poverty estimates that, using statistical techniques, we can make reasonable estimates of changes. We also believe that this is the clearest reading of the statute and is the interpretation intended by Congress, given the sources of data specified in the statute.

D. Dissemination of the Census Bureau Data

Comment: A commenter recommended that the final rule indicate when the child poverty rate will be sent to each State, on the assumption that the Census Bureau could reasonably specify a date by which the child poverty estimates would be available.

Response: The Census Bureau does not have a firm date for issuance of the child poverty estimates. Our plan is to make these data available to the States as soon as they are available from the Census Bureau.

Comment: One commenter made a recommendation for improved communication of the State child poverty rate data. In the NPRM, we proposed to distribute the child poverty rate to the Chief Executive Officer of each State. The commenter recommended that this information also be posted on the DHHS and Census Bureau websites, sent to “Kids Count” organizations in each State, and shared with research and advocacy organizations.

Response: We accept this recommendation in part. The Census Bureau's child poverty data are posted on the Census Bureau's website, and we will post this information on the ACF website. Because this information will thus be readily available, we do not plan to send it to other agencies and organizations.

Section 284.21—What Will We Do if the State's Child Poverty Rate Increased Five Percent or More Over the Two-Year Period?

This new section of the final rule is added for clarity. The content of this new section is derived from § 284.25(d) and (e) and § 284.30(a) of the NPRM and specifies the next steps in the process we will follow if the State's child poverty rate increased five percent or more over the applicable two-year period.

In this section, we provide that, if we determine that the State's child poverty rate increased by five percent or more over the two-year period (based either on the Census Bureau estimates or, if we accept them, the State's independent estimates), we will notify the State that it has 90 days from its receipt of our notification to submit an assessment of the impact of the TANF program(s) in the State on that increase. To provide flexibility, we have added an additional 30 days beyond the 60-day period proposed in the NPRM for States to submit the assessment.

We will also notify those States in which the child poverty rate did not increase five percent or more over the two-year period that no further information or action is required for the applicable two-year period.

Proposed § 284.25—What Information Must the State Provide if the Estimate of the State's Child Poverty Rate Has Increased Five Percent or More Over the Two Year Period? (DELETED IN THE FINAL RULE)

We deleted this section in the final rule.

In the NPRM, we proposed that, if a State's child poverty rate increased by five percent or more, the State could provide information to explain, indicate a subsequent improvement in, or “rebut” this increase. We proposed that the State must submit information on the number of households with children receiving food stamps. The State could also submit school nutrition information and/or an estimate of the State's child poverty rate derived from an independent source. Alternatively, the State could accept the Census Bureau estimate and move to the next step in the process, i.e., to prepare an assessment of the relationship of the TANF program to the increase in child poverty.

We based this proposed step in § 284.25 on our assumption that, despite recent changes, the relationship between the child poverty rate and the food stamp participation rate continued to be a reasonably reliable one. We also proposed this step because we recognized the time lag in receipt of the child poverty estimates. We believed that a State whose child poverty rate increased during the period 1996-1997 should not necessarily be required to assess the relationship to the TANF program and develop a corrective action plan if, in 1998 or 1999, verifiable circumstances indicated that the rate or level of child poverty in the State had improved.

Comments: A majority of commenters strongly objected to this proposal in § 284.25, and almost all urged its deletion. They stated that there is no longer a direct relationship between the numbers of children in poverty and the numbers of children receiving food stamps. Some commenters also pointed out that recent declines in food stamp participation appear to reflect a decline in participation among those eligible for the program as well as a reduction in poverty. Other commenters found that the changes in the food stamp participation rate were related to the implementation of the TANF program, e.g., that State practices, whether intended or unintended, had, indeed, affected the food stamp participation rate.

In addition, commenters expressed concern that using food stamp data to “rebut” increases in the child poverty rate could give States more incentive to ignore administrative problems that could lead to reduced food stamp participation among eligible families with children. Our proposed approach, they believed, could enable, if not encourage, States to avoid taking corrective action to address such administrative problems.

Several commenters also referred to recent national studies that found food stamp declines unexplained by unemployment or other factors and provided specific State data illustrating the lack of relationship in these two data sets over recent years.

A national organization also objected to this provision on the grounds that it presented difficult administrative and reporting burdens for States. This organization and a State TANF agency also objected to the proposed requirement to report on any changes in food stamp policy and procedures, including changes made at the national level, that have affected the food stamp participation rate. (§ 284.25(c)(3) of the NPRM.) They believed that this provision was extremely broad and amorphous and would be both burdensome and costly to States.

Response: After further research and analysis, we agree with these general comments on § 284.25 and have deleted this section from the final rule.

Comment: One national organization recommended that, instead of using food stamp data to explain or “rebut” an increase in child poverty, the final rule should require that States provide data on: (1) The number of families receiving TANF cash payments; (2) the total amount of State spending on TANF cash payments; and (3) the numbers of families and families with children participating in the food stamp program.

Response: We agree that the recommended information may potentially be valuable in assessing the relationship of the TANF program(s) in the State to any increase in child poverty. Rather than using these recommended data as part of a “rebuttal” process, however, we have added the three recommended items to § 284.30(b) as information a State may wish to submit as part of its assessment of the impact of the TANF program on the increase in child poverty.

Comment: Another national organization recommended that the final rule require States to submit data on the employer-reported earnings levels of TANF leavers, e.g., Unemployment Insurance wage record data. (These data are similar to the data provided by States competing for the high performance bonus.) This commenter believed these data would be highly relevant to evaluating the poverty rate in the State. They also recommended that the final rule inform States that submitting multi-year Unemployment Insurance wage record data would be an appropriate and meaningful way to show that an increase in child poverty is not the result of the TANF program.

Response: We agree that the Unemployment Insurance wage record data may be valuable in assessing the possible impact of TANF on State child poverty, and we have added this information in § 284.30(b).

Comment: Two commenters agreed with our recognition of the limited usefulness of the school nutrition information (in assessing the relationship between child poverty and the TANF program(s)) and supported our proposal in the NPRM to make this information optional.

Response: We have continued to make the school nutrition information optional as a part of the State's assessment in § 284.30.

Section 284.30—What Information Must the State Include in Its Assessment of the Impact of the TANF Program(s) in the State on the Increase in Child Poverty?

If a State's child poverty rate increased by five percent or more, we proposed in the NPRM that the State must make an assessment of the impact of the TANF program on the increase in the child poverty rate, covering the same two-year period for which an increase in child poverty was identified. We proposed that the State must submit the assessment, and the information on which the assessment was based, within 60 days. We also listed examples of suggested information and evidence the State might want to include in its assessment.

Comments: Most commenters agreed with our proposal to allow States the flexibility to base their assessment on a wide range of information, including data from other assistance programs, State economic conditions, etc.

Response: We have continued the same general approach in this section as we proposed in the NPRM, but we have made the following changes for additional clarity, specificity, flexibility and in response to comments:

• Retained the requirements that the assessment must cover the same two-calendar-year period as the Census Bureau estimates provided to the State and include the information on which the assessment was based;
• Added, in paragraph (a), that the assessment must directly address the issue of whether the State's child poverty rate increased as a result of the TANF program(s) in the State and include the State's analysis, explanation, and conclusions in relation to this issue to help assure a high quality, focused assessment;
• Provided an expanded list of examples of data and information the State may include in its assessment, including examples of information from Tribal TANF programs; and
• Allowed States 90 days (an additional 30 days beyond the 60 days proposed in the NPRM) to submit the assessment.

Comment: One commenter read the statute as requiring only actions initiated by the State, i.e., the State must specify whether the State's child poverty rate increased by five percent or more and, if so, it must develop a corrective action plan. This commenter objected to the proposed assessment process as beyond the scope of the law.

Response: We believe that section 413(i)(2) of the Act requires the assessment of, and some conclusion regarding, the impact of the TANF program(s) in the State on the child poverty rate before a State moves to develop a corrective action plan. We do not read the statute to require that all States that experience a five percent increase or more in child poverty must develop such a plan—only those where the increase was a result of the TANF program(s) in the State.

However, if a State objects to the assessment process and wishes to conclude that the increase in the child poverty rate is due to, or is the result of, the TANF program(s) in the State, without any analysis or assessment, the State may skip the assessment process and prepare a corrective action plan.

Comment: One commenter objected to our flexible approach to the assessment process on the assumption that asking a State to report on whether its TANF policies contributed to an increase in child poverty put the State in an untenable, conflict-of-interest position. They doubted that any State would self-report any actions that would jeopardize its current practices or policies, particularly those policies related to caseload reduction. They also believed that it would be easy for a State to point to other factors (e.g., economic circumstances) as the primary reason for increases in child poverty.

Response: We recognize the commenter's concern, but we decline to be more prescriptive in our requirements. As we said in the preamble to the NPRM, it is the Department's responsibility to determine whether a State's child poverty rate increased as a result of the TANF program(s) in the State. This is a responsibility we take seriously. Thus, we proposed, in the NPRM, that we would consider not only the information the State submitted in its assessment, but also other information that is readily available, such as State TANF plan provisions, eligibility requirements, benefit levels, TANF expenditures, and other factors.

We also expect that States will take this responsibility seriously. We know States are concerned with the well-being of their citizens, and some are actively addressing issues of poverty, frequently in cooperation with other public and private agencies. We also know that there is much public concern and attention focused on the issue of child poverty by the media, researchers, advocacy organizations, and Congress. We expect that the States will respond to this provision by providing a thorough and reasoned analysis and assessment.

Section 284.35—What Action Will We Take in Response to a State's Assessment and Other Information?

In the NPRM, § 284.35 was titled, “How will the methodology for the Territories differ?”. The section in the final rule regarding the Territories is now numbered § 284.50. We have created a new § 284.35 in the final rule. The content of this new section is taken from paragraphs (b) and (c) of § 284.30 of the NPRM. In the NPRM, these paragraphs proposed that we would review the State's assessment, along with other available information; make a determination whether the child poverty rate increased at least five percent as a result of the TANF program(s) in the State; and notify the State whether a corrective action plan was required.

Comment: One commenter asked that we specify in the final rule how a State (or DHHS) should (or will) attribute any increase in child poverty to the TANF program. They asked that we specify what formula for computation we would use, what criteria we would use, and how States should weight the numerous variables (both TANF and non-TANF related) in the formula that might contribute to such an increase.

Response: We have not specified a computation formula, established criteria, or identified variables for the State assessment process because we do not believe we could, in advance, specify exactly how DHHS will review and evaluate a State's assessment. We believe that a process that relies on analysis, evaluation, and judgment will be more likely to reflect reality rather than a computation formula or weighting of variables. This analysis and judgment will be particularly important when a variety of factors, including the TANF program, may have resulted in the increase in child poverty. Our plan is to work cooperatively with States in reviewing the assessment information and making a final determination on whether a corrective action plan is required. It is not our intention to require a corrective action plan when the TANF program is only a minimal cause of the increase in the State's child poverty rate.

We have made no substantive changes, but we have modified the language in this section for clarity.

Section 284.40—When Is a Corrective Action Plan Due?

In the NPRM, we proposed that a corrective action plan is required only for those States and Territories whose child poverty rate increased by five percent or more as a result of TANF. We also proposed that the State and Territory must submit the plan within 90 days of the date we notify it of our determination that such a relationship exists between the TANF program and the child poverty rate.

Comment: One commenter supported a 90-day timeframe for submitting the corrective action plan.

Response: We have made no changes to this section of the final rule.

Section 284.45—What Are the Contents and Duration of a Corrective Action Plan?

In the NPRM, we proposed in paragraph (a) that the corrective action plan must outline the manner in which the State or the Territory will reduce child poverty in the State and included a description of the actions to be taken by the State or Territory under such a plan.

We proposed in paragraph (b) that the State or Territory must implement the corrective action plan until it determines that the child poverty rate in the State is less than the lowest child poverty rate on the basis of which the State was required to submit the corrective action plan.

Comment: One commenter supported our proposal to allow States the flexibility to design the content of their corrective action plans as States are best able to determine which methods will work best for reducing child poverty within their boundaries. Another commenter recommended that this section be revised to require corrective action plan content specific to affected States, based on Federal site visits and monitoring of States. They believed that, as proposed, this section was weak and ineffective. Also, if a State were allowed to develop its own corrective action plan, it would merely be a “paper exercise.”

Response: The Act does not provide express authority for us to prescribe regulations regarding the content of the corrective action plan. We believe, however, that States will take the requirement to develop a corrective action plan seriously, not only because of concern for the issues affecting the well-being of families and children but also, in part, because of the attention being given to this issue by the media and a wide range of external organizations. We have made no change in this section of the final rule in response to this comment.

Comment: One commenter was concerned that there was no penalty currently assigned to this section of the regulations. The commenter believed that it would be extremely awkward for a State to agree to a corrective action plan only to have a penalty assigned at a later date.

Response: There was no penalty proposed in the NPRM because there was no penalty included in section 413(i) of the Act. Given the limited Federal regulatory and enforcement authority under the TANF program, we have not included a penalty provision. However, we will monitor the State's actions and timelines under the plan. We also expect that interested national, State, and local organizations will monitor State progress in this matter as well.

We want to clarify that the corrective action plan is not intended to be based on or cover the two-year period in which an increase in the child poverty rate was identified. Rather, we anticipate that the State's corrective action plan would include both past efforts and current activities aimed at reducing child poverty.

We have made several changes in this section of the final rule. In § 284.45(a), we added language regarding the inclusion in the corrective action plan of any action steps that will be taken by the Tribes (or consortia of Tribes) under the plan. We also added a requirement in paragraph (b) that the State notify us when it determines that it is no longer required to implement the corrective action plan. Finally, for clarification, we added a definition of the term “lowest child poverty rate” in paragraph (b) and specified that the State must use the methodology in § 284.20 in determining when a corrective action plan is no longer required to be implemented.

We took the definition of the term “lowest child poverty rate” and the following explanatory language from the preamble to the NPRM regarding the duration of the corrective action plan (see 63 FR 50844).

Section 413(i)(4) of the Act requires that the State implement the corrective action plan “until the State determines that the child poverty rate in the State is less than the lowest child poverty rate on the basis of which the State was required to submit the corrective action plan.”

The “lowest child poverty rate” means the five percent threshold above the first year in the two-year comparison period. For example, a State with a 20 percent child poverty rate in the first year of the two-year comparison period would have a five percent threshold of 21 percent and would be required to implement its corrective action plan until its child poverty rate dropped below 21 percent.

By specifying that the State must use the methodology in § 284.20 in determining the duration of the corrective action plan, we intend to clarify that the State may use either the Census Bureau data or an independent estimate of the child poverty rate.

Section 284.50—What Information Will We Use To Determine the Child Poverty Rate in Each Territory? (§ 284.35 of the NPRM)

The Census Bureau produces annual estimates of the child poverty rate in each of the 50 States and the District of Columbia. However, the Census Bureau does not develop poverty estimates for the Territories. Therefore, in § 284.35 of the NPRM, we proposed that each Territory must, annually, beginning in 1998 (for calendar year 1996), submit to ACF certain food stamp or other data on which we would calculate a child poverty rate. We also proposed a process similar to the proposed State process for determining whether the child poverty rate increased by five percent or more between the applicable years. Finally, we specified the actions and information we would require if the child poverty rate increased by five percent or more as a result of the TANF program.

“Territories” are defined, for the purposes of this part, as American Samoa, Guam, Puerto Rico, and the United States Virgin Islands. At the present time, this part applies to all of these jurisdictions except American Samoa, which has not applied to operate a TANF program. When it does so, the provisions of this part will apply to it as well.

We received no comments on § 284.35 of the NPRM or how we proposed to determine a child poverty rate for the Territories. We did, however, receive comments critical of our proposed use of State food stamp data in rebutting the increase in child poverty (in § 284.25). These comments caused us to re-evaluate our use of food stamp data as the basis for calculating a child poverty rate for the Territories.

During the development of the final rule, we had numerous discussions with the Census Bureau and others in an attempt to identify reliable child poverty data for the Territories, but we were unsuccessful. Therefore, we have revised § 284.50 to postpone, temporarily, the determination of a child poverty rate for these jurisdictions. However, we are committed to applying the provisions of section 413(i) to both States and Territories. We specify, in paragraph (a) that our intent is to apply the same requirements and procedures to the Territories as to the States. We specify in § 284.50(b) that we will estimate the number of children in poverty in these jurisdictions when reliable data are available.

V. Regulatory Impact Analyses

A. Executive Order 12866

Executive Order 12866 requires that regulations be drafted to ensure that they are consistent with the priorities and principles set forth in the Executive Order.

The Department has determined that this rule is consistent with these priorities and principles. This rulemaking implements statutory authority based on broad consultation and coordination. It also reflects our response to comments received on the NPRM issued on September 23, 1998.

The Executive Order encourages agencies, as appropriate, to provide the public with meaningful participation in the regulatory process. As described elsewhere in the preamble, ACF consulted with State and local officials, their representative organizations, researchers, a broad range of technical and interest group representatives, and others to obtain their views prior to the publication of the NPRM.

To a considerable degree, the NPRM reflected the information provided by, and the recommendations of, the groups with whom we consulted. We also carefully considered and have accepted and/or responded to the comments received in response to the NPRM.

This final rule also reflects the intent of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 to achieve a balance between granting States the flexibility they need to develop and operate effective and responsive programs and ensuring that they meet the objectives of the statute. The limited scope of this regulation is also consistent with the provisions of the statute and Administration policy as articulated in Executive Order 12866 and its Regulatory Reinvention Initiatives.

The Department has determined that this rule is significant under the Executive Order. The Office of Management and Budget has reviewed this rule. This rule is not a major rule under the Small Business Regulatory Enforcement Fairness Act of 1996.

We have estimated the maximum annualized Paperwork Reduction Act costs to be approximately $454,118. This is clearly an upper limit on what the costs would be if all States were required to respond to all requirements. Thus, as discussed in section D below, this figure is an over-estimate of the expected costs.

In assessing the potential costs and benefits—both quantitative and qualitative—of these final regulations, the Secretary has determined that the benefits of these regulations justify the costs. The Secretary has also determined that this regulatory action does not unduly interfere with State, local, and Tribal governments in the exercise of their governmental functions.

B. Regulatory Flexibility Analysis

The Regulatory Flexibility Act (5 U.S.C. 603, 605) requires the Federal government to anticipate and reduce the impact of rules and paperwork requirements on small businesses and other small entities. Small entities are defined in the Regulatory Flexibility Act to include small businesses, small non-profit organizations, and small governmental agencies. This rule will immediately affect the 50 States and the District of Columbia. It will affect the Territories in the future, when reliable Census Bureau data on child poverty in the Territories are available. Therefore, the Secretary certifies that this rule will not have a significant impact on small entities.

C. Assessment of the Impact on Family Well-Being

We certify that we have made an assessment of this rule's impact on the well-being of families, as required under section 654 of The Treasury and General Government Appropriations Act of 1999. The overall aim of the TANF program is to strengthen the economic and social stability of families. The purpose of this rule is to monitor annual estimates of child poverty in the States (and, in the future, the Territories); assess the impact of the TANF program(s) on an increase in child poverty of five percent or more; and require the development of a corrective action plan, if indicated. We believe that the well-being of families will be enhanced by this rule.

D. Paperwork Reduction Act

This rule contains three information collection requirements. These requirements were reviewed and approved by the Office of Management and Budget (OMB) at the NPRM stage on December 2, 1998, under the Paperwork Reduction Act of 1995 (OMB control number 0970-0186). This data collection approval expires on November 30, 2001. An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB control number.

To the extent possible, this rule relies on existing data sources. We will obtain data on child poverty from the Census Bureau for the 50 States and the District of Columbia. We have postponed implementing the provisions of this final rule for the Territories until reliable Census Bureau data on child poverty in the Territories are available.

The three information collection activities in the final rule are: (1) As an optional provision, a State or Territory may provide an alternative estimate of the child poverty rate (§ 284.20(c)); (2) a State or Territory may be required to conduct and submit an assessment of the impact of the TANF program(s) in the State or Territory on the increase in the child poverty rate (§ 284.30 and § 284.50); and (3) a State or Territory may be required to submit a corrective action plan (§ 284.40, § 284.45 and § 284.50). These information collection requirements were approved at the NPRM level. We received no comments on the burden as proposed in the NPRM.

The annual burden estimates include any time involved compiling and abstracting information, analyzing and evaluating information, assembling materials necessary to provide the requested information, obtaining clearance, and transmitting the information.

Prior to the development of this estimate, we researched the burden estimates for similar OMB-approved data collections in our inventory and those pending OMB approval and consulted with knowledgeable Federal officials.

The 50 States and the District of Columbia are the immediate potential respondents to the information collection requirements in this rule. These jurisdictions may, at their option, submit an estimate of child poverty from an independent source. They may also be required to submit an assessment and a corrective action plan. We will not implement these information collection activities for the Territories until we have reliable child poverty data for these jurisdictions, but we have included Puerto Rico, Guam, and the Virgin Islands in the burden calculation as they will be respondents in the future.

We have increased the estimated total annual burden from 15,240 hours in the NPRM to 15,552 hours in the final rule. This change reflects the elimination of the requirement that the Territories provide data for us to compute an estimate of their child poverty rate (§ 284.35 of the NPRM); the elimination of the requirement that States provide food stamp data and other data to explain or rebut an increase in the State's child poverty rate (§ 284.25 of the NPRM); the addition of 8 hours per respondent for the optional submission of data on a State's child poverty rate from an independent source (§ 284.20(c) of the final rule); and an increase in the estimated burden hours for a State to develop and submit an assessment of the impact of TANF on the child poverty rate from 80 hours per respondent in the NPRM to 120 hours in the final rule (§ 284.30 of the final rule).

The annual burden estimates for each of the three data collections are:

We have estimated the burden hours for each information collection activity in part 284 as though they applied to all jurisdictions for ease of discussion and public review. This is clearly an over-estimate. We do not expect that all States (or Territories) will opt to provide an alternate estimate of child poverty derived from an independent source. We expect that no more than a few States (or Territories) will experience an increase of five percent or more in their child poverty rates and will need to submit an assessment in relation to the TANF program; and only a few States (or Territories) will be required to submit a corrective action plan.

We estimate the annualized cost of the hour burden to be $454,118. Again, this is an over-estimate. It is based on an estimated average hourly wage of $29.20 (including fringe benefits, overhead, and general and administrative costs) for the State staff performing the work multiplied by the estimated 15,552 burden hours.

We expect that no capital/start-up costs and operation/maintenance costs will be required as a result of a State or Territory's implementation of this part. No systems modifications should be required and much of the information that States may submit as a part of their assessment is pre-existing information or available from other State executive branch agencies or research sources. Therefore, we do not anticipate any significant costs beyond the annualized cost of the hour burden noted above.

We considered all comments by the public in:

• Evaluating whether the collections are necessary for the proper performance of our functions, including whether the information will have practical utility;
• Evaluating the accuracy of our estimate of the burden of the collections of information, including the validity of the methodology and assumptions used, and the frequency of collection;
• Enhancing the quality, usefulness, and clarity of the information to be collected; and
• Minimizing the burden of the collection of information on those who are to respond, including through the use of appropriate automated, electronic, mechanical, or other technology, e.g., the electronic submission of responses.

E. Unfunded Mandates Reform Act of 1995

Section 202 of the Unfunded Mandates Reform Act of 1995 requires that a covered agency prepare a budgetary impact statement before promulgating a rule that includes any Federal mandate that may result in the expenditure by State, local, and Tribal governments, in the aggregate, or by the private sector, of $100 million or more in any one year.

If a covered agency must prepare a budgetary impact statement, section 205 further requires that it select the most cost-effective and least burdensome alternative that achieves the objectives of the rule and is consistent with the statutory requirements. In addition, section 203 requires a plan for informing and advising any small government that may be significantly or uniquely impacted by the rule.

We have determined that this rule will not result in the expenditure, in the aggregate, by State, Territorial, local, and Tribal governments, or by the private sector, of more than $100 million in any one year. Accordingly, we have not prepared a budgetary impact statement, specifically addressed the regulatory alternatives considered, or prepared a plan for informing and advising any significantly or uniquely impacted small government.

F. Congressional Review

This rule is not a “major” rule as defined in 5 U.S.C., Chapter 8.

G. Executive Order 13132

On August 4, 1999, the President issued Executive Order 13132, “Federalism.” The purposes of the Order are: “to guarantee the division of governmental responsibilities between the national government and the States that was intended by the Framers of the Constitution, to ensure that the principles of federalism established by the Framers guide the executive departments and agencies in the formulation and implementation of policies, and to further the policies of the Unfunded Mandates Reform Act * * *.”

We certify that this final rule does not have a substantial direct effect on States, on the relationship between the Federal government and the States, or on the distribution of power and responsibilities among the various levels of government. The final rule does not pre-empt State law and does not impose unfunded mandates.

This rule does not contain regulatory policies with federalism implications which would require specific consultations with State and local elected officials. However, during the development of the NPRM, we held two types of consultations. First, we raised issues related to this provision in the general TANF consultation meetings with representatives of State and local governments; nonprofit, advocacy, and community organizations; foundations; and others. Second, we held consultations focussed specifically on this provision on May 30, 1997, and September 4, 1997, with national organizations representing State and local elected officials; technical, statistical, and policy experts; and research, advocacy, and public interest organizations that focus on poverty and child well-being.

We sent invitations to the May 30 meeting (along with a list of policy issues proposed for discussion) to, among others: The National Governors' Association (NGA), the National Conference of State Legislatures (NCSL), the National Association of Counties, the National League of Cities, and the United States Conference of Mayors. In addition to these groups, invitations to the September 4 meeting were also sent to the National Black Caucus of State Legislators and the National Organization of Black County Officials. Based on our records, representatives of NGA and NCSL attended both meetings.

List of Subjects in 45 CFR Part 284

• Grant programs—Social programs
• Public Assistance programs; Poverty; Reporting and recordkeeping requirements

(Catalogue of Federal Domestic Assistance Programs: 93.558 Temporary Assistance for Needy Families; 93.559 Federal Loans for State Welfare Programs; 93.594 Tribal Work Grants; and 93.595 Welfare Reform Research, Evaluations and National Studies.)

For the reasons set forth in the preamble, we are amending 45 CFR Chapter II by adding part 284 to read as follows:

PART 284—METHODOLOGY FOR DETERMINING WHETHER AN INCREASE IN A STATE OR TERRITORY'S CHILD POVERTY RATE IS THE RESULT OF THE TANF PROGRAM

Authority: 42 U.S.C. 613(i)