Alicia H. Munnell, Center for Retirement Research, Boston College

Each year, the Social Security actuaries project the system’s financial outlook over the next 75 years, and these projections are published in the annual Trustees Report.  Since 1989, these Reports have included data on future benefits as a percent of pre-retirement earnings – commonly called replacement rates – for workers at different places on the income scale.  That is, the Reports did include these data until this year, when the trustees decided to delete them.

The deletion of replacement rate data is the culmination of a concerted effort by critics who argue that people will have plenty of money in retirement.  With respect to Social Security, these critics contend that the reported replacement rates grossly understate the program’s contribution to retirement income.  In 2013 testimony before the House Ways and Means Committee, one critic reported that according to his calculations Social Security replaced 69 percent of pre-retirement earnings rather than the 40 percent reported in the Trustees Reports (Biggs 2013).

So what’s going on here?  The critics’ main argument is over the measure of pre-retirement earnings used in calculating replacement rates – the Social Security actuaries’ use career average earnings adjusted for economy-wide wage growth (a measure recommended by the Organization for Economic Cooperation and Development), while the critics prefer the last five years of earnings before retirement.  Strikingly, though, a recent study by the actuaries showed that these two types of measures produced very similar numbers.  Using a random sample of 200,000 workers claiming benefits in 2011 with an average retirement age of 63.75, the actuaries found a 38.8 percent replacement rate using the career average approach and a 39.2 percent rate using the last five years of significant earnings.

How do critics get such high replacement rates?  They use earnings in the last five years including years of zero earnings.  In 2011, 15 percent of people claiming benefits had no earnings in the 5 years immediately prior to claiming.  For these workers, the replacement rate would be infinite!!  It makes no sense to include zero years in the denominator of the replacement rate calculation.

So no rational case can be made for deleting replacement rate numbers from the Trustees Report, and doing so has serious policy implications.  If the only numbers provided to policymakers are dollar amounts rising over time – without any reference to the earnings these benefits are replacing – they will think that slowing the rate of increase would do little harm.  But slowing the growth in inflation-adjusted benefits reduces the percent of earnings replaced.

For example, consider the gradual increase in the Full Retirement Age from 65 to 67.  The Table below shows its impact on replacement rates (for an individual claiming benefits at age 65) between 2003 – the year in which the FRA began to rise – and 2030 – after the FRA increase is fully phased in.  The replacement rate for the medium earner drops from 42 percent to 36 percent, with commensurate cuts for workers at all levels on the earnings scale.

Table. Social Security Replacement Rates at Age 65, 2003 and 2030

Earnings level

Replacement rate















Source: U.S. Social Security Administration, Office of the Chief Actuary; and

In short, excluding replacement rate data from the widely-read Trustees Report is unhelpful for both analysts and policymakers.  These data will be provided by the Office of the Chief Actuary in the form of recurring actuarial notes.  And we have put a copy of the missing data on our website.  But this is a silly way to do business.

The whole purpose of the attack on Social Security replacement rates is an attempt to provide a rationale for cutting benefits.  But if Social Security replaces less, then future workers must depend on what is now a fairly wobbly 401(k) system for more.  Without replacement rate numbers, policymakers will have no idea what they are doing to the retirement security of future workers as they consider alternative Social Security provisions.

This is not inside baseball.  This is important.

Posted on: September 2, 2014

One Comment

  1. September 3, 2014 at 12:32 pm - Reply

    This is an important subject.
    This is an important subject. As Prof. Munnell references my work on replacement rates, I’ve put up a response over at AEI’s blog:

Leave A Comment

How can we help you?

Media Inquiries
Sponsorship Opportunities

Stay up-to-date on the latest research & policy updates.

Subscribe to our newsletter