Federal income-security programs provide cash or in-kind benefits to individuals. Some, such as Food Stamps, Supplemental Security Income, Temporary Assistance for Needy Families, and the earned income tax credit, are means-tested; others, including unemployment compensation and civil service retirement and disability payments, are not tied to recipients' income or assets.

Retirement and disability programs—including military retirement—constitute the largest portion of federal spending in function 600, accounting for about one-third of the category's mandatory spending. Refundable tax credits make up 17 percent of mandatory spending for 2007, a growth of 61 percent from 2002. Food and nutrition assistance is the next-largest component, making up about 16 percent of mandatory outlays in recent years. Unemployment compensation, which in 2002 and 2003 made up nearly 20 percent of mandatory spending, has fallen back toward the 10 percent mark of the middle to late 1990s. Housing assistance accounts for about 70 percent of discretionary income-security spending.

The Congressional Budget Office estimates that outlays in function 600 will total $364 billion in 2007 and that about $56 billion of that amount will be discretionary spending. Since 2002, growth in spending for the function has averaged about 3 percent annually. That growth is the net effect of legislation that enhanced refundable tax credits (which are recorded as outlays), a decline in unemployment compensation, and slowing growth in the Food Stamp program.

The Pension Benefit Guaranty Corporation (PBGC) is a federal agency that insures participants in private employers' defined-benefit pension plans against the loss of specified benefits if their plans are terminated without sufficient assets. Private employers are not required to provide a pension for their workers, but if they do, they must follow rules specified in the Employee Retirement Income Security Act (ERISA) for most major aspects of the plan's operation (including minimum standards for participation, accrual of benefits, vesting, and funding). If a plan is terminated with insufficient assets to pay promised benefits, PBGC takes over the plan's assets and liabilities (up to an annual per-participant limit). It uses the assets of the terminated plans along with insurance premiums from ongoing plans to make monthly annuity payments to qualified retirees and their survivors. At the end of 2006, PBGC reported a deficit of about $18 billion— indicating that its assets were about $18 billion less than the present value of the benefits that it owed to workers and retirees in underfunded plans that were terminated or whose termination the agency viewed as "probable."

PBGC's insurance premium for a single-employer plan consists of three parts: a fixed annual payment ($31 in 2007) for each participant (worker or retiree) in the plan; for an underfunded plan, a variable payment equal to $9 for each $1,000 by which the plan is underfunded; and for a plan terminated after January 2006, a $1,250 payment for each participant in each of the first three years following the sponsor's emergence from bankruptcy. In 2006, offsetting receipts (a credit against direct spending) from the fixed portion of the premium totaled about $984 million; and from the variable portion, about $595 million.

Under one alternative, this option would increase collections from the fixed-rate premium by 15 percent. The increase could occur by either increasing the current charge from $31 per participant to $35 or by changing the assessment base to some measure of insured benefits and setting the premium to a level yielding 15 percent more collections. This component of the option would increase offsetting receipts by more than $160 million in 2008 and by about $900 million over five years.

Under a second alternative, which could be pursued singly or in combination with the first, this option would increase the variable-rate premium by one-third, to $12 per $1,000 of underfunding. This change would increase offsetting receipts by about $300 million in 2008 and by $2.3 billion over five years. Under a third alternative, which also could be pursued on its own or along with the first, this option would apply the higher variable-rate premium only to sponsors with a credit rating below single-A. This alternative would increase offsetting receipts by more than $200 million in 2008 and by $1.6 billion over five years.

A principal advantage of increasing premiums is that doing so could improve PBGC's long-term financial condition. Raising premiums for riskier plans would also align premiums more closely with the risk they pose to PBGC. Currently, premiums increase only with underfunding, even though other factors, such as the financial condition of the sponsors and the share of plans' assets allocated to risky securities, also increase the risk to PBGC. By raising the cost of maintaining underfunded or riskier plans, this option would provide an added incentive to employers to more fully fund their plans. Moreover, the current per-participant charge may constitute a disadvantage for new firms with a disproportionate share of employees who have accumulated few pension benefits. An advantage to changing the assessment base for the fixed-rate premium rather than increasing the per-participant charge is that doing so would more directly relate the fixed-rate premium to the level of insured benefits.

A disadvantage of this option is that the increases in premiums would not necessarily be well targeted to plans that PBGC eventually took over because charges would still be relying on the amount of underfunding in a plan and its bond rating, which are not perfectly related to the probability that the plan will be terminated. In addition, raising the insurance premiums for underfunded plans may not be enough to improve PBGC's long-term financial condition. Finally, for financially weak employers, higher premiums could increase the risk that they would terminate their plans.

The government's major retirement plans for civilian employees—the Civil Service Retirement System (CSRS) and the Federal Employees Retirement System (FERS)—provide initial benefits (those provided before cost-of-living adjustments are applied) that are based on the average of an employee's highest earnings over three consecutive years. In 2008, outlays for pension benefits under the two programs are projected to total $62.8 billion.

This option would use a five-year average instead of a three-year average to compute benefits for workers who retire under CSRS and FERS after September 30, 2007. As a result, initial pensions would be about 3 percent smaller for most new civilian retirees, saving the federal government $42 million in 2008 and a total of $1.2 billion over five years. The average new CSRS retiree would receive about $1,250 less in 2008 and $6,530 less over five years than under current law. By comparison, the average new FERS retiree would receive $420 less in 2006 and $2,190 less over five years.

Under an alternative approach, a four-year average would be used for CSRS and FERS. Such action would yield savings of $20 million in 2008 and $580 million over five years. The average new CSRS retiree would receive $600 less in 2008 and $3,140 less over five years while the average new FERS retiree would receive $200 less in 2008 and $1,060 less over five years.

One rationale for using a longer average is that it would better align federal practices with those in the private sector, which commonly uses five-year averages to calculate a worker's base pension. The change in formula also would encourage some federal employees to work longer in order to boost their pensions. That incentive could help the government retain experienced personnel.

A rationale against this option is that cutting pension benefits would reduce the attractiveness of the government's civilian compensation package. In addition, this option would reduce benefits more under CSRS than under FERS because CSRS provides a bigger defined-benefit pension than FERS. Federal employees under FERS also participate in Social Security and receive government contributions to the 401(k)-like Thrift Savings Plan, while those employees under CSRS do not.

Federal pension payments to 4.2 million retired federal workers (both civilian and military) and their survivors are projected to be $104 billion in 2007. For the same year, pension payments to about 516,000 veterans and their survivors are projected to be $3.2 billion, and compensation to 3 million disabled veterans and their survivors is projected to be $30 billion. In addition, veterans' readjustment benefits (including education and vocational rehabilitation benefits) for about 513,000 veterans and their survivors are projected to be $3 billion. All those benefits are indexed to the increase in the consumer price index for urban wage earners and clerical workers (CPI-W), but the size of the adjustment varies among programs, as does the age at which benefits are payable.

• Pensions paid to former federal workers under the Civil Service Retirement System (CSRS) are subject to annual cost-of-living adjustments (COLAs) that provide complete protection against increases in the CPI-W. Pensions paid under the newer Federal Employees Retirement System (FERS) are fully protected only when that increase is less than 2 percent a year. If the percentage increase in the CPI-W is between 2 percent and 3 percent, FERS annuitants receive a COLA of 2 percent. If the CPI-W increase exceeds 3 percent, their COLA is the percentage increase in the CPI-W minus 1 percentage point. In addition, FERS annuitants receive COLAs only at ages 62 and above—except for those retiring on disability. Both CSRS and FERS participants can generally start receiving pension benefits at age 60 with 20 years of service or at age 62 with five years of service. However, participants with 30 years of service can retire at earlier ages.
  
  
• Pensions paid to military retirees hired before August 1, 1986, qualify for full COLAs. Military personnel who enter service after that date face a choice: They can elect to stay under the old system and receive a full COLA, or they can opt to receive a $30,000 bonus at their 15th year of service and receive reduced annual COLAs that equal the percentage increase in the CPI-W less 1 percentage point. Under the latter arrangement, when those retirees turn 62, they receive a one-time adjustment or "catch-up" that restores their annuity to what it would have been had the full COLA been paid. After age 62, they continue to receive the reduced COLA. However, to date, most military personnel have opted to forgo the bonus at 15 years of service and thus remain eligible for the full COLA. Military personnel who retire from the active-duty force are eligible to receive their pension benefits after completing 20 years of service without any accompanying age requirements. (Reservists are not eligible for retirement annuities until they reach 60.) Personnel with fewer than 20 years of service generally are not eligible for any benefits unless they qualify for retirement on disability.
  
  
• Veterans' benefit programs—including disability compensation and death benefits for survivors, vocational rehabilitation, pensions, and education benefits for veterans and their survivors—all receive full COLAs. Disability benefits are payable to veterans who have certified disabilities, with the amount of the payment depending upon the severity of the disability. Veterans also are eligible for means-tested pension benefits.

This option would base all cost-of-living adjustments for federal civilian and military retirees and veterans' benefits on the increase in the chained CPI, an alternative measure of inflation developed by the Bureau of Labor Statistics. The Congressional Budget Office estimates that the chained CPI is likely to grow 0.3 percentage points more slowly per year than the standard CPI.

Under this option, annual COLAs would equal the increase in the chained CPI for CSRS annuitants, pre-1986 military retirees, and veterans. Comparable adjustments would be made for FERS annuitants, who would receive the full COLA when the increase in the chained CPI was less than 2 percent a year; a COLA of 2 percent when the increase in the chained CPI was between 2 percent and 3 percent; and a COLA 1 percentage point below that increase when it exceeded 3 percent. Military retirees under the new system would receive a reduced COLA equal to the percentage growth in the chained CPI minus 1 percent.

Those changes would decrease mandatory outlays by $332 million in 2008 and by $6.3 billion over the 2008-2012 period. On average, a CSRS retirement annuitant would receive about $1,230 less over five years than under current law, and a FERS retirement annuitant would receive about $300 less. The average military retiree would receive roughly $1,250 less over five years relative to current law, whereas veterans' disability compensation would fall by about $410 and veterans' pensions by about $390. (Using the chained CPI for all federal benefit programs that are indexed for inflation would reduce outlays by $35.4 billion over the 2008-2012 period and by nearly $140 billion through 2017.)

A rationale for limiting COLAs is that many analysts believe the CPI-W overstates increases in the cost of living and that using the alternative measure would reduce federal outlays while ensuring that benefits did not fall any lower in real terms than they were when the recipients became eligible for the programs. (For more details, see the discussion in Option 650-1.) In addition, federal pension plans offer greater inflation protection than most private pension plans do, and COLAs are becoming increasingly scarce in the private sector. According to a 2001 survey, fewer than 15 percent of private-sector plans gave annuitants formal annual COLAs; another 25 percent made cost-of-living adjustments on an ad hoc basis. More than 60 percent of plans had made no adjustments during the previous 10 years. Moreover, even with reduced COLAs, many federal and military annuitants would still fare better than other retirees because they are covered by the comprehensive Federal Employees Health Benefits program or have access to military treatment facilities, and Tricare, the military's health plan. Veterans also have access to federal facilities.

Various arguments against limiting COLAs also could be made. Using the alternative measure of inflation could be more difficult to implement because the chained CPI is subject to revisions. In addition, that measure might understate changes in the cost of living for retirees, whose expenditure patterns differ from those of the general population. In that case, limiting COLAs could allow the benefits received by both current and future retirees to decline over time in real terms. CSRS annuitants would be particularly affected because they are most dependent on their pensions (and may have stayed with CSRS on the basis of the understanding that they would enjoy the current COLA protection against inflation.) Moreover, because current and prospective employees take into account the relatively large retirement benefits offered by the government when comparing alternative wage and benefit packages, reducing federal retirement benefits could affect the government's ability to recruit and retain a highly qualified workforce. Finally, because military personnel can retire at earlier ages and receive an immediate pension after just 20 years of service, lower COLAs for them would have bigger effects over longer periods. And, reducing veterans' education benefits might reduce the number of veterans who could afford to continue their education.

Today, most federal workers covered by the Federal Employees Retirement System (FERS) can direct up to $15,500 of their salary to the Thrift Savings Plan (TSP), which is similar to a 401(k) plan. (Employees who are 50 and older are able to make an additional "catchup" contribution of up to $5,000.) The Internal Revenue Service sets the contribution limit and adjusts it annually. Previously, the limits capped workers' contributions at 15 percent of their salary, but that percentage cap was removed in January 2006. Under FERS, federal agencies automatically contribute an amount equal to 1 percent of an employee's salary to the TSP; agencies also match the first 3 percent of workers' voluntary contributions to the TSP dollar for dollar and match the next 2 percent of contributions at 50 cents on the dollar. Thus, although those employees can save higher shares of their earnings in the TSP, they receive the maximum government match by contributing just 5 percent. (Federal workers covered by the Civil Service Retirement System, the older federal plan, can contribute up to $15,500 of their salary to the TSP, but they receive no government match.)

This option would restructure the TSP contribution schedule so that the government made the full 5 percent match only when employees contribute 10 percent of their salary. Specifically, federal agencies would continue to automatically contribute an amount equal to 1 percent of employees' earnings and match the first 2 percent of voluntary contributions dollar for dollar (a maximum match of 2 percent). Contributions ranging from 3 percent to 10 percent would be matched at 25 cents perdollar (a maximum match of another 2 percent). That restructuring would save $359 million in 2008 and $2.1 billion over the 2008-2012 period.

A justification for changing the government's matching schedule is that it would bring federal practices more in line with those of the private sector, which usually provides lower matches and no automatic contributions to defined-contribution plans. For example, according to the Bureau of Labor Statistics, the most prevalent practice among medium-sized and large private firms is to match employees' contributions of up to 6 percent of pay at 50 cents on the dollar. This option would also give some FERS employees, especially those now contributing 5 percent of their earnings, incentive to set aside more in the TSP and thus have more savings available when they retire. Furthermore, restructuring matching contributions might reduce the disparity that currently exists between the government's two major retirement systems. In most cases, the benefits that an employee receives under FERS—which include Social Security and the TSP—cost the government more than the benefits that the same employee would receive under the Civil Service Retirement System.

There are several rationales against implementing the option. First, a lower government match on smaller contributions could reduce some workers' incentive to participate in the TSP or to continue contributing at their current rates. Second, the government might be faulted for saving money at the expense of those employees who are least likely to contribute a higher percentage of their earnings to the TSP—young workers and others with relatively low pay. Third, changing the TSP could be considered unfair because one factor that affected many people's decision to accept employment with the government or to switch from the Civil Service Retirement System to FERS was their assumption that TSP benefits would remain the same.

The Trade Adjustment Assistance (TAA) program offers income-replacement benefits, training, and related services to workers who lose a job as a result of competition from imports or a shift of production to another country. To obtain assistance, affected workers must first petition the Secretary of Labor for certification and then meet other eligibility requirements. Cash benefits are available to certified workers who receive training, but only after they have exhausted their unemployment insurance benefits. Legislation enacted in 2002 expanded eligibility for the program and provided displaced workers with a refundable tax credit of 65 percent of their health insurance premiums.

Ending the TAA program by issuing no new certifications in 2008 and thereafter would reduce federal outlays by about $500 million in 2008 and by $4.5 billion through 2012. Affected workers could still apply for benefits under the Workforce Investment Act of 1998, which authorizes a broad range of employment and training services for displaced workers regardless of the cause of their job loss. (This option would also increase revenues because it would eliminate the refundable tax credit for a portion of their health insurance costs that TAA beneficiaries qualify for. Revenues would be about $20 million higher in 2008, the Joint Committee on Taxation estimates, and $150 million higher through 2012.)

A rationale for this option is that such a change would help ensure that federal programs offered more-uniform assistance to workers who were permanently displaced as a result of changing economic conditions. Because the Workforce Investment Act provides cash benefits only under limited circumstances and does not provide a subsidy for health insurance premiums, workers who lose a job because of foreign competition or as a result of a shift in production to another country are generally treated more generously than are workers who are displaced for other reasons.

An argument against this option is that eliminating TAA benefits could cause economic problems for some workers who have been unemployed for a long time and who otherwise would have received such benefits. Another way of securing more-equal treatment for displaced workers, regardless of the reason for their job loss, would be to expand benefits for displaced workers who are not currently eligible for the TAA program.

Most low-income tenants who receive federal rental assistance are aided through the Housing Choice Voucher program, the low-rent Public Housing program, or project-based assistance programs (which designate privately owned government-subsidized units for low-income tenants). Administered by the Department of Housing and Urban Development (HUD), those programs usually require that tenants pay 30 percent of their monthly gross household income (after certain adjustments) for rent; the federal government subsidizes the difference between that amount and the maximum allowable rent. In 2006, the Congressional Budget Office estimates, the average federal expenditure for all of HUD's rental housing programs combined was roughly $7,070 per assisted household. That amount includes both housing subsidies and fees paid to agencies that administer the programs.

This option would increase tenants' rent contributions over a five-year period, raising them from 30 percent of adjusted gross income to 35 percent. Provided that federal appropriations are lowered to offset those higher contributions by tenants, savings in outlays would total $370 million in 2008 and $7.3 billion over five years, including $3.3 billion for the Housing Choice Voucher program, $1.6 billion for the Public Housing program, and $2.4 billion for project-based assistance programs.

An argument for this option is that its effects on tenants could be cushioned by encouraging states to make up some or all of the decrease in federal support. States currently contribute no funds to federal rental assistance programs, even though such programs generate substantial local benefits, including improvements in the quality of the housing stock and in the general welfare of assisted tenants.

An argument against the option is that some states might not increase their spending to compensate for the reduction in federal assistance. As a result, housing costs could rise for current recipients of aid. For those with the very lowest income, even a modest increase in rent could be difficult to manage. Moreover, by increasing the percentage of their income that tenants are required to pay toward rent, the option would reduce their incentive to work.

Recipients of federal housing assistance typically live either in subsidized-housing projects or in rental units of their own choosing found on the open market. Financial support for the second type of assistance usually comes in the form of vouchers—specifically, from the Housing Choice Voucher program. Administered by the Department of Housing and Urban Development, that program pays the difference between a tenant's contribution (usually 30 percent of his or her monthly adjusted gross income) and rent (which is determined by local rent levels).

Both the local payment standard and the federal subsidy vary according to the type of unit in which a given tenant lives. Generally, an individual in a one-person household may choose an apartment with up to one bedroom. Recipients in larger households may rent larger units.

This option would link the rent subsidy for new applicants from one-person households to the cost of an efficiency apartment rather than a one-bedroom unit. (The change would also apply to any single person receiving assistance who moved to another subsidized unit.) Provided that federal appropriations are adjusted accordingly, the option would save $13 million in outlays next year and $392 million through 2012.

A rationale for this option is that an efficiency unit should provide adequate space for someone who lives alone. A potential drawback is that renters in some areas might have difficulty finding an efficiency apartment and, under the new rule, might have to spend a larger percentage of their income for a one-bedroom unit.

Under the Food Stamp program, applicants must meet eligibility requirements to receive a monthly benefit. In general, among other conditions, household income must be at or below 130 percent of the federal poverty line, and countable assets must be less than $2,000.

Once eligibility for the program has been determined, the amount of the benefit is calculated. A household is expected to contribute 30 percent of its net income (gross income minus deductions for certain expenses) toward food expenditures. The Department of Agriculture has calculated the monthly cost of a so-called Thrifty Food Plan for households of various sizes. The Food Stamp benefit equals the amount by which the monthly cost of the Thrifty Food Plan exceeds 30 percent of a given household's net monthly income. A minimum benefit has been set for one- and two-person households: If the calculated benefit is less than $10, the Food Stamp benefit is $10.

This option would eliminate Food Stamp benefits for one- and two-person households whose calculated benefit was less than $10 a month. The change would reduce outlays by $35 million in 2008 and by $315 million over five years.

A rationale for this option is that it would reserve Food Stamp benefits for recipients who had the greatest calculated need. An argument against the option is that eliminating Food Stamps for households that currently are eligible for benefits of less than $10 a month might discourage those households from applying for the program if their financial situation worsened. If the option did discourage such applications, it would lessen the extent to which the program achieved its goal of aiding low-income households.

The School Lunch Program and the School Breakfast Program provide funds that enable participating schools to offer subsidized meals to students. In general, participating schools offer free meals to students whose household income is at or below 130 percent of the federal poverty line, reduced-price meals to students whose household income is above 130 percent but at or below 185 percent of the federal poverty line, and full-price meals to students whose household income is above 185 percent of the poverty line.

The subsidy rate per meal does not vary with the cost that a given school incurs as a result of providing the lunch or breakfast—it depends solely on the household income of the student who receives the meal. For the 2006-2007 school year, federal cash subsidies total $2.40 per free lunch and $1.31 per free breakfast served; $2.00 per reduced-price lunch and $1.01 per reduced-price breakfast served; and $0.23 per full-price lunch and $0.24 per full-price breakfast served. (Schools in Alaska and Hawaii and schools that have large numbers of students who participate in the free- and reduced-price-meal programs receive an additional subsidy.) Although each school sets the prices it charges students for reduced- and full-price meals, the reduced-price lunch may not cost more than $0.40, and the reduced-price breakfast may not cost more than $0.30.

This option, which would yield net reductions in outlays of $75 million in 2008 and more than $2.5 billion over five years, would eliminate the breakfast and lunch subsidy for full-price meals for students whose household income is above 350 percent of the poverty line, beginning in July 2008. At the same time, it would increase the subsidy for a reduced-price meal (both breakfast and lunch) by $0.20.

A rationale for this option is that there is no clear justification for subsidizing meals for students who are not from low-income households. An argument against the option is that if a participating school has been using funds from the full-price subsidy to offset the overall costs of administering its breakfast and lunch programs, it might decide to raise meal prices for students from higher-income households, or it might drop out of the program altogether. The latter outcome would mean that students who were eligible for free or reduced-price meals would no longer receive them.

The federal Supplemental Security Income (SSI) program provides monthly cash payments—based on uniform, nationwide eligibility rules—to low-income elderly and disabled people. In addition, many states provide supplemental payments. Because SSI is a means-tested program, recipients' non-SSI income can reduce their SSI benefits, subject to certain exclusions. For unearned income (which is mostly Social Security benefits), $20 a month is excluded from the benefit calculation; above that amount, SSI benefits are reduced dollar for dollar. To encourage SSI recipients to work, the program allows a larger exclusion for earned income.

This option would lower the exclusion for unearned income from $20 a month to $15, reducing outlays by $110 million in 2008 and by $720 million over five years.

A rationale for this option is that a program designed to ensure a minimum standard of living for its recipients does not need to provide a higher standard for those people who happen to have unearned income (generally, Social Security benefits). An argument against the option is that reducing the monthly exclusion by $5 would decrease by as much as $60 the income of the roughly 2.8 million low-income people (approximately 40 percent of all federal SSI recipients) who otherwise would benefit in 2008 to a greater extent from the exclusion.

The federal Supplemental Security Income (SSI) program makes cash payments to low-income elderly and disabled people on the basis of uniform, nationwide eligibility rules. In addition, many states provide supplemental payments to program recipients. In 2007, children will receive approximately $7 billion, or about one-fifth of total benefits.

Unlike other means-tested benefits, SSI payments for each additional child do not decline as the number of SSI recipients in a family increases. For instance, in 2007, a family that includes one child who qualifies for SSI benefits can expect to receive up to $623 a month if the family's income (excluding SSI benefits) is under the cap for the maximum benefit. If the family includes other eligible children, it can receive another $623 a month for each additional child. (A child's benefit is based on the presence of a severe disability and on the family's income and resources. Neither the type of disability nor participation by other family members in the SSI program is considered.)

This option would create a sliding scale for SSI disability benefits so that a family would get incrementally fewer benefits per child as the number of children in the family who qualified for SSI increased. If the option was implemented in 2009 (to allow the Social Security Administration, which administers the SSI program, to gather data on multiple SSI recipients in individual families), outlays would drop by $80 million in 2009 and by $605 million between 2009 and 2012.

Recommended by the National Commission on Childhood Disability in 1995, the sliding scale that this option presents would keep the maximum benefit for one child at the level currently allowed by law. However, benefits for each additional child in the same family would be correspondingly reduced. If the sliding scale was applied in 2007, the first child in a family who qualified for the maximum benefit would continue to receive $623 a month. But the second child would get $389, and the third would receive $332. Benefits would continue to decrease for additional children in the same family. As with current SSI benefits, the payments would be adjusted each year to reflect changes in the consumer price index.

Proponents of a sliding scale argue that the resulting reductions in benefits would reflect economies of scale that generally affect the cost of living for families who have more than one child. Furthermore, the high medical costs that disabled children often incur, which would not be subject to economies of scale, would continue to be covered because SSI participants are generally eligible for Medicaid.

An argument against this option is that children with disabilities sometimes have unique needs (such as housing modifications and specialized equipment) that may not be covered by Medicaid. If SSI benefits were reduced, some families might be unable to meet those needs.

The federal Supplemental Security Income (SSI) program makes monthly cash payments to low-income elderly and disabled people. The Social Security Administration (SSA), which administers the program, sometimes pays recipients more than it later determines they should have received. According to a report issued by the General Accounting Office (now the Government Accountability Office), the complexity of the rules that govern the SSI program is a primary reason for the overpayments.¹

After discovering an overpayment, the SSA may reduce the recipient's subsequent monthly benefit to recover the excess amount. Under current rules, however, the maximum that the SSA may deduct from a recipient's monthly payment is the lesser of two amounts: the recipient's entire monthly SSI benefit or 10 percent of the recipient's total monthly income (minus certain exclusions). Thus, the SSA may deduct no more than 10 percent of the monthly SSI benefit of a recipient who has no other source of income. Moreover, the Commissioner of Social Security may lower the rate of recovery or waive collection of an overpayment altogether if it is determined that doing so would support the purposes of the SSI program.

This option would remove the ceiling on the amount of overpayments that the SSA could recover from monthly SSI payments but retain the commissioner's discretionary authority to reduce or waive the required amount. Removing the 10 percent ceiling would increase the amount collected from overpayments—and thereby reduce net outlays for benefits—by $70 million in 2008 and by $490 million over the 2008-2012 period. (Removing the ceiling would increase administrative costs by about $35 million to $45 million each year; however, those costs are subject to the appropriation process and are not included in the amounts shown in the table.)

An argument in support of this option is that removing the ceiling would improve the federal government's ability to recover money paid to recipients erroneously. Moreover, retention of the commissioner's discretionary authority would lessen the chances that such recoveries would result in undue hardship for SSI recipients.

An argument against the option is that SSI recipients generally have low income and few, if any, financial assets. For recipients who have no other income, even a 10 percent reduction in their SSI payments might be difficult to manage. The current ceiling allows affected recipients to pay the amount they owe in small increments, whichlimits the reduction they must make in their current spending.

The Child Care and Development Block Grant, which provides money to states to subsidize the child care expenses of low-income families, is funded through a combination of discretionary appropriations and a capped entitlement. Created in 1990, the program was subsequently modified and reauthorized through 2002 as part of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. Between 2002 and 2005, the capped entitlement—which included annual increases through 2002 under the 1996 law—was held at its 2002 level of $2.7 billion per year and was not adjusted for inflation. That part of the block grant was recently increased, as part of the Deficit Reduction Act of 2005, to $2.9 billion per year through 2010.

This option would increase the 2008 authorization for the entitlement portion of the block grant to adjust for inflation since 2006 and would index that amount thereafter. That change would boost federal spending by $123 million in 2008 and by $1.5 billion through 2012.

A rationale for indexing the entitlement portion of the block grant is that it would maintain low-income families' access to subsidized child care. That access, in turn, would increase the incentive to work for some low-income parents, making it easier for them not only to enter the job market but also to stay employed. Increased participation in paid child care might also improve children's well-being, potentially decreasing their behavioral problems while increasing their social skills and their readiness to enter school.

An argument against this option is that many low-income parents have access to informal, or unpaid, care (from a relative, for example). In those cases, increases in child care subsidies might simply result in those parents' shifting from unpaid to paid care. Furthermore, there is little evidence about the effects on children of informal (as opposed to paid) care.