Raising the Limit on Social Security Covered Earnings

10/06/2011
Featured in print Bulletin on Aging & Health

The U.S. Social Security system faces a substantial long-term funding gap. One of the most commonly suggested remedies is to increase the level of earnings subject to Social Security taxes. Currently, only earnings up to $106,800 are taxed, though workers pay Medicare taxes on earnings above this level. Raising or eliminating the cap on earnings subject to Social Security taxes would generate additional revenues for the system.

A less well-understood effect of raising the limit on covered earnings is that it would increase Social Security benefits for workers, since benefits are based on workers' covered earnings over their best thirty-five years. This effect would tend to offset, at least in part, the additional revenues from the earnings limit increase.

In The Growth in Social Security Benefits Among the Retirement Age Population from Increases in the Cap on Covered Earnings (NBER Working Paper 16501), researchers Alan Gustman, Thomas Steinmeier, and Nahid Tabatabai examine this issue. The authors make use of the actual changes in the limit on covered earnings over the past several decades to ask how the taxes and benefits of more recent cohorts would have differed if these workers had been subject to the limits faced by earlier cohorts.

While the limit on covered earnings is now adjusted automatically each year to match the growth in average wages, in the past the limit was often unchanged for long periods of time and then subject to dramatic increases. The percent of male workers with earnings above the limit rose from 5 percent in 1940 to nearly 50 percent by 1965, before declining to around 10 percent by 1990 and stabilizing at this level. These dramatic changes mean that workers from different birth cohorts with the same actual earnings profiles could have very different covered earnings histories.

The authors use data from the Health and Retirement Study. Their basic approach is to take younger cohorts (born 1948-1953) and recompute their Social Security taxes and benefits applying the limits on covered earnings faced by cohorts born twelve and twenty-four years earlier. All other factors, such as workers' real earnings, are held constant.

The authors' key finding is that the higher limits on covered earnings faced by younger cohorts increased their Social Security taxes and benefits by 3.7 percent and 1.5 percent, respectively, relative to the levels they would have experienced with the limits in place twelve years earlier (and by 10.7 and 5.3 percent relative to the limits in place twenty-four years earlier). Approximately 25 to 30 percent of the additional tax revenue raised was diverted to pay for the benefit increases.

As expected, the increases in taxes and benefits are highly concentrated among men in the top quartile of earnings - their taxes and benefits increased by 12.7 percent and 3.9 percent relative to the level of twelve years earlier (and by 26.7 and 10.2 percent relative to the level of twenty-four years earlier). Increases for men in lower earnings quartiles and for women were smaller.

The authors conclude by pointing out that an increase in the limit on covered earnings is different from an across-the-board increase in the payroll tax rate in two ways. First, raising the earnings limit restricts the tax increase to those with the highest earnings. Second, this approach suffers from a "leaky bucket" problem, as part of the new tax revenue raised will be needed to cover the increase in benefits. While in theory the benefit formula could be changed to eliminate the benefit increase, some may be reluctant to do so because they feel this would violate the insurance principle underlying the program. This paper suggests that the tax increase resulting from a rise in the covered earnings limit has been three to four times the size of the benefit increase.


This research was supported by a grant from the U.S. Social Security Administration through the Michigan Retirement Research Center.