Leave the CPI and COLAs Out of the Budget Talks

December 6, 2012

Download Fiscal Fact No. 345: Leave the CPI and COLAs Out of the Budget Talks

Press reports indicate that a key feature of the House Republicans’ recent deficit reduction proposal is to switch from the regular consumer price index (CPI) to the chained CPI for adjusting various federal tax and spending formulas for inflation. The rationale for this proposal is that, over the last dozen years, the chained CPI has risen about a half percent less each year than the regular CPI. Thus, the switch would, among other things, slow the adjustment of the income tax brackets for inflation and reduce the annual cost of living increases (COLAs) for Social Security retirees. Both would be bad policy.

Tax rate increases in disguise

The proposed switch would tamper with the indexing feature of the income tax, the last remaining unaltered piece of Economic Recovery Tax Act signed by Ronald Reagan in 1981. Reducing the adjustment of the income tax brackets for inflation would push people more quickly over time from one tax bracket to the next higher one as incomes grow. Pushing more people from the 15 percent bracket to the 25 percent bracket, or from the 28 percent bracket to the 33 percent bracket, is as much a marginal tax rate increase as any other type of tax rate hike. It would raise marginal tax rates faster over time than under current law and is simply an attempt to disguise a tax rate hike as something else.

The progressive tax rate structure is supposed to make people with higher incomes pay more tax than people with lower incomes in any given year. It should not enable the government to automatically take a rising share of the national income without a vote in Congress as Americans’ incomes grow over time with advances in technology and labor productivity.

To the extent that the regular CPI adjustment may slightly exceed inflation in some cases, it also protects taxpayers against a sliver of “real bracket creep” from real wage gains. That protection promotes job growth and the health of the economy. With real bracket creep, the bulk of the workforce could eventually end up in the top tax brackets. Even middle- and low-income earners would be facing European-level income tax rates and perhaps European-level unemployment rates.

COLAs not the cause of Social Security deficits

COLAs that are slightly overstated by the much maligned CPI are not the source of Social Security's looming deficits, and fixing the CPI and COLAs will not save Social Security from impending insolvency.

There are two real culprits leading to the projected Social Security deficits. One is an adverse demographic shift as Baby Boomers retire, life expectancy rises, and the population ages. Currently there are 2.8 workers paying into the system for each retiree drawing benefits. That ratio will fall to 2.5 workers per retiree by 2020, 2.0 workers per retiree by 2055, and to 1.9 workers per retiree by 2090. The decline in the ratio stems in part from the impending retirement of the post-World War II baby boom generation, which will be followed into the work force by the baby bust generation.

Exacerbating the population shift is a continuing increase in life expectancy. According to the 2012 Social Security Trustees Report, life expectancy at age 65 in 2011 was 18.7 years for men and 20.7 years for women.[1] The Report projects those figure to reach 21.5 and 23.2 years, respectively, by 2050, and 23.2 and 24.8 years, respectively, by 2090.[2] A larger group of retirees will live and draw benefits 20 to 25 percent longer than they do today.

The other source of the deficits is the Social Security initial benefit formula. The initial benefit formula sets the benefit check that a retiree receives when he or she begins to draw benefits. The benefit is based on a worker's earnings history—the income on which he or she paid Social Security taxes over the years. Those yearly earnings are adjusted for wage growth, averaged, and subjected to a complex formula that determines each person's initial monthly benefit check.

The initial benefit formula is designed to keep benefits growing in line with income, generation after generation, no matter how much income grows over time. The "replacement rates" (initial benefits upon retirement as a percent of pre-retirement income) are kept nearly constant across all future generations—at about 55 percent for low wage workers, 41 percent for average wage workers, and 27 percent for upper income workers (see Table 1). They drop only when the normal retirement age is raised, as will happen when it goes from age 66 to age 67 for people turning age 62 between 2017 and 2022.

Per capita real income (over and above inflation) is projected to rise by between 160 percent and 170 percent—more than double, nearly triple—over the next 78 years. Consequently, per capita real benefits at the normal retirement age are projected to rise, by roughly 140 percent to 150 percent, under current benefit formula rules between 2012 and 2090, held down a bit by a one-year rise in the normal retirement age. A single professional worker retiring at 66 in 2012 could start out with a benefit of $29,902; a professional married couple with as much as $59,804. A similar worker retiring at age 67 in 2090 could get an initial benefit of about $73,850, and a couple about $147,700, in today's dollars if real wages rise as projected (see Table 1).

There is no legitimate policy reason to give future, and richer, retirees benefits that are equal to three times today’s average family income when they can well afford to save for themselves and when the outlays would drive the system and the government into financial crisis.

The initial benefit formula is independent of the subsequent cost of living adjustments that protect the initial benefit from subsequent inflation. Consequently, changes in the COLAs have little impact on the initial benefits that future generations will get.[3] Any savings die with each cohort, and new entrants come in with the same unadjusted initial benefits. Switching from the regular CPI to the chained CPI would trim the COLA less than a half percent a year. At the end of a twenty year average retirement, it would trim the benefit by about 10 percent, or about 5 percent on average over the life of the recipient. Since the savings die with the recipient, that is the most the system’s outlays would be trimmed.

Table 1. Estimated Pre-retirement Income and Real Benefit Amounts of Retired Single Workers Upon Retirement

at Normal Retirement Age* With Various Pre-retirement Earnings Levels** Based on Intermediate Assumptions

Benefits, constant 2012 dollars

Percent of career-average earnings

Career-average earnings, constant 2012 dollars***

Year Attaining Age 65

Low Earnings

Average Earnings

Maximum Earnings

Low Earnings

Average Earnings

Maximum Earnings

Low Earnings

Average Earnings

Maximum Earnings









































*Normal retirement age at which full benefits are payable is currently 66. It is scheduled to rise to 67 in stages. The table shows benefits for an individual retiring at age 66 in 2012; two months later in 2020; and one year later, at age 67, in 2055 and 2090.
**Low earnings equal 45 percent of average earnings. Average earnings assume worker earned national average covered earnings each year of working life. Maximum earnings assume worker earned the SSA contribu­tion and benefit base (maximum covered earnings) each year of working life. Source: 2012 Annual Report of the Board of Trustees of the Federal Old Age and Survivors Insurance and Disability Insurance Trust Funds, Table V.C7.
***Career-average earnings, wage indexed to the year prior to retirement, expressed in 2012 dollars.

Instead of attacking the COLAs of current retirees, it would make more sense to slow the growth of real initial benefits for future retirees by amending the benefit formula and/or raising the normal retirement age another year. The current benefit calculation uses wage growth to adjust workers’ earnings and the “bend points” or “brackets” in the current benefit formula. Switching to price growth for these adjustments instead of wage growth (“price indexing” instead of “wage indexing”) would slow the growth of real benefits. Instead of growing between 140 percent and 150 percent, real benefits would grow roughly between 50 percent and 70 percent. A bit over half of the projected real benefit increases would be forestalled. That could save ten times the money as trimming COLA's a half percent per year.

Other methods of using price indexing have been offered. One, called progressive price indexing, would allow some wage-adjusted growth at the bottom "bracket" or "bend point" in the benefit formula while adjusting the top end of the benefit formula by prices, to raise lower income benefits relative to upper income benefits. There are an infinite numbers of ways of bringing the OASDI retirement system into balance or surplus, prospectively, with ample warning and without cutting current COLAs for current retirees.

People who have tried repeatedly to patch up the system are always surprised that it keeps slipping back into deficit, but there is no mystery as to why. It is not possible to pay larger real benefits for a rising number of years of retirement with a declining number of workers per beneficiary without substantial tax rate increases. Furthermore, every effort to patch up the system has reinforced the adverse demographic trends that render the system irreparable. Each payroll tax increase has discouraged employment, saving, and income. People are young before they are old. It makes no sense to lower their after-tax incomes, employment opportunities, and ability to save while they are of working age in order to give them higher transfer payments after they retire.

[1] Social Security Administration, 2012 OASDI Trustees Report, http://www.ssa.gov/oact/tr/2012/index.html.

[2] Id.

[3] The CPI has a minor effect on the initial benefit calculation if the worker delays taking benefits past age 62.



About the Author

Related Research