Mandatory Spending

Multiple Budget Functions

Use an Alternative Measure of Inflation to Index Social Security and Other Mandatory Programs

CBO periodically issues a compendium of policy options (called Options for Reducing the Deficit) covering a broad range of issues, as well as separate reports that include options for changing federal tax and spending policies in particular areas. This option appears in one of those publications. The options are derived from many sources and reflect a range of possibilities. For each option, CBO presents an estimate of its effects on the budget but makes no recommendations. Inclusion or exclusion of any particular option does not imply an endorsement or rejection by CBO.

Billions of Dollars 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2019-
Change in Outlays  
  Social Security 0 -2.0 -4.8 -7.9 -11.2 -14.6 -18.1 -21.6 -25.1 -28.8 -25.9 -134.1
  Other benefit programs with COLAsa 0 -0.5 -1.3 -2.2 -2.9 -3.4 -4.4 -5.2 -6.1 -7.2 -6.9 -33.3
  Effects on SNAP from interactions with COLA programsb 0 0.1 0.1 0.1 0.2 0.3 0.3 0.3 0.4 0.5 0.5 2.2
  Health programs 0 -0.3 -1.2 -2.0 -2.7 -3.6 -4.6 -5.5 -6.6 -7.8 -6.3 -34.3
  Other federal spendingc 0 * -0.1 -0.2 -0.2 -0.3 -0.4 -0.5 -0.6 -0.8 -0.5 -3.2
    Total 0 -2.9 -7.3 -12.1 -16.8 -21.7 -27.2 -32.6 -38.0 -44.1 -39.1 -202.7
Change in Revenuesd 0 * * * * * * * * * -0.1 -0.2
Decrease (-) in the Deficit 0 -2.9 -7.2 -12.1 -16.8 -21.6 -27.2 -32.6 -38.0 -44.0 -39.1 -202.4

Sources: Congressional Budget Office; staff of the Joint Committee on Taxation.
This option would take effect in January 2020.
COLA = cost-of-living adjustment; SNAP = Supplemental Nutrition Assistance Program; * = between -$50 million and $50 million.
a. Other benefit programs with COLAs include civil service retirement, military retirement, Supplemental Security Income, veterans' pensions and compensation, and other retirement programs whose COLAs are linked directly to those for Social Security or civil service retirement.
b. The policy change would reduce payments from other federal programs to people who also receive benefits from SNAP. Because SNAP benefits are based on a formula that considers such income, a decrease in those other payments would lead to an increase in SNAP benefits.
c. Other federal spending includes changes to benefits and various aspects (eligibility thresholds, funding levels, and payment rates, for instance) of other federal programs, such as those providing Pell grants and student loans, SNAP, child nutrition programs, and programs (other than health programs) linked to the federal poverty guidelines. (The changes in spending on SNAP included here are those besides the changes in benefits that result from interactions with COLA programs.)
d. The effects on revenues reflect the reduction in marketplace subsidies for health insurance premiums and slightly higher enrollment in employment-based coverage under the option.


Cost-of-living adjustments for Social Security (COLAs) and many other parameters of federal programs are indexed to increases in traditional measures of the consumer price index (CPI). The CPI measures overall inflation and is calculated by the Bureau of Labor Statistics (BLS). In addition to the traditional measures of the CPI, that agency computes another measure of inflation—the chained CPI—designed to account for changes in spending patterns and to eliminate several types of statistical biases that exist in the traditional CPI measures. (Nonetheless, the chained CPI does not resolve all statistical issues with traditional CPI measures.) Under current law, beginning in 2018, the chained CPI would be used for indexing most parameters of the tax system, including the individual income tax brackets.


Beginning in 2020, this option would use the chained CPI for indexing COLAs for Social Security and for indexing parameters of other programs. The chained CPI has grown an average of about 0.25 percentage points more slowly per year since 2001 than the traditional CPI measures have, and the Congressional Budget Office expects that gap to persist. Therefore, the option would reduce federal spending, and savings would grow each year as the effects of the change compounded.

Effects on the Budget

Outlays would be reduced by $203 billion through 2028, CBO estimates, and the net effect on the deficit would be about the same. The budgetary effects of this option would stem from a reduction in the average benefits that eligible people receive through a number of federal programs, and, to a lesser extent, from a reduction in eligibility for certain programs. (The small revenue effects estimated here are the net result of two largely offsetting factors. First, the option would reduce marketplace subsidies for health insurance premiums. Because those subsidies are structured as refundable tax credits, a portion of the reduction in subsidies translates into higher tax liabilities for recipients, meaning higher revenues. Second, slightly higher enrollment in employment-based coverage under the option would mean that a larger share of compensation would be made in the form of nontaxable health benefits, which would result in less taxable compensation for employees, and, therefore, less revenues.)

The CPI affects COLAs for Social Security and the pensions that the government pays to retired federal civilian employees and military personnel, as well as veterans' pensions and veterans' disability compensation. In most of those programs, the policy change would not alter benefits when people are first eligible to receive them, either now or in the future, but it would reduce their benefits in later years because the annual COLAs would be smaller, on average. The effect would be greater the longer people received benefits (that is, the more years of reduced COLAs they experienced). Therefore, the effect would ultimately be especially large for the oldest beneficiaries as well as for some disabled beneficiaries and military retirees, who generally become eligible for annuities before age 62 and thus can receive COLAs for a longer period.

To obtain the estimates for the effects of the option on COLAs, CBO reduced payments for beneficiaries after the first year of receipt by the difference between the traditional CPI and the chained CPI in each year. For example, in the case of COLAs for Social Security, CBO estimates that about 63 million people would be affected by the benefit reductions in 2020, experiencing an average benefit reduction of about 0.25 percent relative to current law. By 2028, the average reduction in monthly benefits for those people is projected to be 2.2 percent relative to current law.

By affecting program parameters, growth in the CPI also affects spending for Supplemental Security Income, Medicare, Medicaid, the health insurance marketplaces established under the Affordable Care Act, Pell grants, student loans, the Supplemental Nutrition Assistance Program (SNAP), child nutrition programs, and other programs. The index is used to calculate various eligibility thresholds, payment rates, and other factors that could affect the number of people eligible for those programs and the benefits people receive. For some programs, such as Medicaid, budgetary savings stem from the reduction in the number of people eligible for those programs and from the reduction in the average federal spending on each eligible person. For other programs, such as Medicare, savings from this option stem largely from reductions in the updates to prices that the federal government would pay.

For SNAP, the option would lead to higher spending as a result of two opposing effects. On the one hand, the policy change would lead to a reduction in SNAP benefits. The amount of those benefits is based on beneficiaries' total income minus allowable deductions, such as costs associated with housing and child care, and the value of some of those deductions in each year is linked to the CPI. Lower deductions would lead to lower SNAP benefits. On the other hand, a reduction in payments from other federal programs as a result of the option would reduce beneficiaries' income, leading to higher SNAP benefits. Because that second effect is larger, the option would increase SNAP benefits, on net.

The uncertainty in the estimate of budgetary savings from this option stems from differences between the projected traditional CPI and chained CPI. Historically, that gap has varied widely. For example, in calendar year 2005, the chained CPI growth was 0.51 percentage points slower than the CPI for all urban consumers, and in calendar year 2008, growth was 0.12 percentage points faster.

Other Effects

One argument for switching to the chained CPI in Social Security and other federal programs is that the chained CPI is generally viewed as a more accurate measure of overall inflation than the traditional CPI measures, for two main reasons. First, the chained CPI more fully accounts for how people tend to respond to price changes. Consumers often lessen the effect of inflation on their standard of living by buying fewer goods or services that have risen in price and by buying more goods or services that have not risen in price or have risen less. Measures of inflation that do not account for such substitution overstate growth in the cost of living—a problem known as substitution bias. BLS's procedures for calculating the traditional CPI measures account for some types of substitution, but the chained CPI more fully incorporates the effects of changing buying patterns.

A second reason to believe that the chained CPI is a better measure of inflation is that it is largely free of a problem known as small-sample bias. That bias, which is significant in the traditional CPI measures, occurs when certain statistical methods are applied to price data for only a limited number of items in the economy.

One argument against using the chained CPI, and thereby reducing COLAs in Social Security and other federal retirement programs, is that the chained CPI might not accurately measure the growth in prices that Social Security beneficiaries and other retirees face. The elderly tend to spend a larger percentage of their income on items whose prices can rise especially quickly, such as health care. (However, determining how rising health care prices affect the cost of living is problematic because accurately accounting for changes in the quality of health care is challenging.) The possibility that the cost of living may grow faster for the elderly than for the rest of the population is of particular concern because Social Security and pension benefits are the main source of income for many retirees.

Another argument against this option is that a reduction in COLAs would ultimately have larger effects on the oldest beneficiaries and on the disabled beneficiaries who received benefits for a longer period. For example, if benefits were adjusted every year by 0.25 percentage points less than the increase in the traditional CPI measures, Social Security beneficiaries who claimed benefits at age 62 would face a reduction in retirement benefits at age 75 of about 3 percent compared with what they would receive under current law, and a reduction at age 95 of about 8 percent. To protect vulnerable people, lawmakers might choose to reduce COLAs only for beneficiaries whose income or benefits were greater than specified amounts. Doing so, however, would reduce the budgetary savings from the option.

Finally, policymakers might prefer to maintain current law because they want benefits to grow faster than the cost of living so that beneficiaries would share in overall economic growth. An alternative approach would be to link benefits to wages or gross domestic product. Because those measures generally grow faster than inflation, such a change would increase outlays.