Retirements Are At Risk

Mobile Share Email Facebook Twitter LinkedIn
Headshot of Alicia H. Munnell

is a columnist for MarketWatch and director of the Center for Retirement Research at Boston College.

The “all fine” crusaders are selective in their use of numbers.

Keeping up with the campaign that the retirement system is fine and people will have plenty of money in retirement is a full-time job.  So far, critics have pointed to three main statistics to bolster their case.

  1. The vast majority of today’s workers – about 80 percent – are covered by an employer-sponsored pension plan on their current job.
  2. Retirement income from employer-sponsored pensions, particularly 401(k)s, is much higher than previously thought because the widely used Current Population Survey understates such income.
  3. Social Security benefits replace more of workers’ pre-retirement earnings than previously thought because the widely used measure of Social Security replacement rates is flawed.

In the first two cases, the critique contains a kernel of fact but the information is not applicable to the argument.  In the last case – Social Security replacement rates – I don’t know what they are thinking.  Let’s take a look at each point.    

  1. Critics state that 78 percent of workers have employer pension coverage on their current job.  This figure is based on the Labor Department’s National Compensation Survey.  But it incorporates both public and private sector workers, focuses on access rather than participation, and ignores part-time workers.  Public sector workers have near universal pension coverage, so the coverage problem is about the private sector.  Individuals who fail to participate in an offered plan are a concern because they need to participate in order to build retirement savings.  And, finally, there is no obvious rationale for excluding part-time workers from a measure of pension coverage; when they retire, they will also need savings to supplement their retirement income.  Taking these factors into account, the actual private sector pension participation rate drops to 48 percent in the National Compensation Survey.  And other surveys show somewhat lower participation rates.
  2. Critics have expressed concern that the Census Bureau’s popular Consumer Population Survey (CPS) grossly understates the income that retirees draw from their 401(k)s and IRAs.  They maintain that if such income were factored in, retirees would look much better off.  On the first point, the critics are correct – the CPS, apparently due to current flaws in the survey questions, does fail to capture most of the income generated by 401(k)s/IRAs – compared to either the Federal Reserve’s Survey of Consumer Finances or the income reported to the IRS.  However, a careful analysis of these different data sources shows that the impact of this flaw is largely limited to the top quintile of the income distribution, because low- and middle-income retirees have very little in 401(k)/IRA assets.  The CPS should be fixed and Census is currently devising changes to address the problem, but the bottom line is that today’s CPS measure of retiree income is largely accurate for the bulk of households.
  3. Critics do not question the dollar amount of Social Security benefits, but rather the measure of pre-retirement earnings to which it is compared (known as a replacement rate).  The standard measure, recommended by the OECD, uses wage-indexed career-average earnings and shows that the medium earner retiring today at age 65 receives about 40 percent of pre-retirement income from Social Security.  And these replacement rates will drop as Social Security’s so-called “Full Retirement Age” continues to rise.

    The critics use two unusual alternative measures of pre-retirement income to contend that Social Security benefits are much more generous.  The first defines a worker’s pre-retirement earnings as average earnings in the last five years before retirement, including years with zero earnings.  An analysis by the Social Security actuaries found that 14 percent of individuals had no earnings in the five years before they claimed, which means that their replacement rate would be infinite!  If only years with significant earnings are included, the actuaries calculate that the replacement rate is virtually identical to their standard measure.  

    The second measure uses price-indexed career-average earnings, thereby excluding real growth in wages and, thus, living standards.  Interestingly, though, an analysis by the Social Security actuaries shows that this price-indexed measure produces a replacement rate for the medium earner that is about 44 percent, only modestly higher than the standard measure.

For those of you following the debate, I hope this info is helpful.