What Would Be the Effect on the Deficit of Using the Chained CPI to Index Benefit Programs and the Tax Code?

Posted by
Emily Stern
on
April 18, 2013

This morning my colleague Jeffrey Kling testified before the Subcommittee on Social Security of the House Committee on Ways and Means and described the budgetary effects of switching to the chained consumer price index (CPI) to index benefit programs and the tax code. Cost-of-living adjustments (COLAs) for Social Security benefits and other parameters of many federal programs and the tax code are currently indexed to increases in the traditional CPI, a measure of overall inflation calculated by the Bureau of Labor Statistics (BLS). According to many analysts, however, the CPI overstates increases in the cost of living because it does not fully account for the fact that consumers generally adjust their spending patterns as some prices change relative to other prices and because of a statistical bias related to the limited amount of price data that BLS can collect.

One option for lawmakers would be to link federal benefit programs and tax provisions to another measure of inflation—the chained CPI—that is designed to account fully for changes in spending patterns and that does not have the same statistical bias. The chained CPI grows more slowly than the traditional CPI does: an average of about 0.25 percentage points more slowly per year over the past decade. As a result, using that measure to index benefit programs would reduce federal spending for Social Security, federal employees’ pensions, Medicare, Medicaid, and various other programs. In addition, indexing tax provisions with the chained CPI would increase revenues.

Specifically, CBO and the staff of the Joint Committee on Taxation estimate that switching to the chained CPI-U on a governmentwide basis starting in calendar year 2014 would reduce the deficit by a total of $340 billion over the next 10 years. Such a change would decrease federal spending on mandatory programs (direct spending) by $216 billion and increase federal revenues by $124 billion over the fiscal year 2014–2023 period. (This estimate was first published by CBO on March 1. It is not an estimate of the proposal in the President’s budget for fiscal year 2014, which would use the chained CPI for a more limited set of programs. CBO is currently reviewing that and other proposals in the President’s budget.)

A little more than half of the reduction in spending would be for Social Security. According to CBO’s analysis, using the chained CPI-U for annual COLAs would reduce outlays for Social Security (relative to CBO’s current-law baseline) by $1.6 billion in 2014. Those savings would grow each year, reaching $24.8 billion in 2023, and would total $127 billion over the 2014–2023 period. CBO projects that Social Security recipients would face an average benefit reduction of 0.25 percent in 2014 (about $3 per person per month) and approximately 2 percent in 2023 (roughly $30 per person per month). That estimated average reduction in 2023 reflects larger percentage cuts (of up to 2.5 percent) for people who are already receiving benefits today or will become eligible for them shortly (and who thus would have experienced smaller COLAs for nearly a decade by 2023) and smaller cuts for people who will become eligible for benefits later (and thus would have experienced smaller COLAs for a shorter period of time in 2023). By 2033, outlays for Social Security would be 3 percent lower than they would be under current law, or 6.0 percent of gross domestic product (GDP) rather than 6.2 percent. As a result, the gap between Social Security’s outlays and tax revenues in that year would shrink by about one-sixth, to 1.0 percent of GDP.

Switching to the chained CPI-U governmentwide would also lower benefits in other programs that apply automatic COLAs, including civil service and military retirement, Supplemental Security Income, and veterans’ programs. In addition, the change would reduce federal spending for Medicaid, Medicare, higher education assistance, and nutrition programs, among other mandatory programs. In the case of certain means-tested programs, such as Medicaid and nutrition assistance, those reductions would occur in part because using the chained CPI U to make annual adjustments to the federal poverty guidelines would decrease eligibility for those programs.

The impact of using the chained CPI-U would vary among participants in the affected programs. Where the index was used to inflate a benefit or payment level, such as with Social Security, all program participants would receive a lower benefit than they would under current law. Where the chained CPI-U was used to inflate a threshold, such as the federal poverty guidelines, there would be a large effect on participants who lost eligibility for certain benefits but no effect on other program participants.

In the case of Medicare, for example, switching to the chained CPI-U would affect both payment rates and thresholds for means-tested elements of the program. CBO estimates that such a policy would reduce net Medicare spending per beneficiary by an average of roughly $3 per month in 2023. Of that amount, about $2 per beneficiary, on average, would reflect reductions in payments to providers and plans for services furnished to beneficiaries; those reductions would affect payments for services furnished to most beneficiaries. The remaining reduction of roughly $1 per beneficiary, on average, would stem from two factors: First, roughly half a million beneficiaries would see their premiums for Parts B and D of Medicare increase by up to $125 per month because they would become subject to higher premiums on the basis of their income. Second, Medicare spending would be reduced by an average of about $300 a month for approximately 100,000 beneficiaries who would qualify for the Part D low-income subsidy program (LIS) under current law, because those beneficiaries would receive less generous or no LIS subsidies for Part D premiums or cost sharing under the policy.

Emily Stern is an analyst in CBO’s Budget Analysis Division.