Want a more comfortable retirement? Put it off

Michael Maiello | Sep 20, 2018

Sections Economics Finance

Retirement planning is one of the most complex financial conundrums facing American households. It requires setting savings rates that anticipate unpredictable investment returns, estimated spending needs, and an uncertain lifespan. No one aspect of it is ever fully in the planner’s control, which means that choosing the right retirement date can be difficult and fraught with uncertainty.

The answer may be simply to put off retirement as long as possible. Delaying retirement by as little as three months can have a significant effect on retirement living standards, according to Gila Bronshtein of Cornerstone Research, Jason Scott of Financial Engines, Stanford’s John B. Shoven, and George Mason University’s Sita N. Slavov

Bronshtein, Scott, Shoven, and Slavov used national demographic and income averages to model households facing retirement decisions, then verified their conclusions by examining the experiences of actual households represented in the University of Michigan’s Health and Retirement Study, a cooperative project with the US’s National Institute on Aging that has surveyed the US population aged 50 and older since 1992.

They find that “delaying retirement by three to six months has the same impact on the retirement standard of living as saving an additional one-percentage point of labor earnings for 30 years. The relative power of saving more is even lower if the decision to increase saving is made later in the work life.”

The study makes two key assumptions. First, workers annuitize their savings at retirement. The younger they are, the more expensive the annuity will be, so delaying it by a year makes a huge difference in costs. This assumption facilitates comparison between social security and private retirement savings, but it does not affect the researchers’ broad conclusions; even for workers who do not annuitize their savings, retiring later reduces the period that must be funded by private savings. 

The second assumption is that retirees take social security at retirement. Although social security does not need to be taken at retirement, previous research from Shoven, Slavov, and Harvard’s David A. Wise has found that most Americans do claim it at the time they retire. Social-security benefits are more generous when taken later in life, and, since they make up the bulk of most Americans’ retirement savings, should be deferred if possible.

For a couple whose primary earner fits the researchers’ “base case”—someone who starts saving for retirement at 36 and contributes 6 percent of earnings to a 401(k) with a 3 percent employer match, with wage growth and asset returns equal to inflation—the real standard of living in retirement increases by nearly 8 percent when the primary earner retires at 67 instead of 66.

The research explores alternative strategies for improving retirement income, including making use of lower-cost investment portfolios and starting saving at a younger age. While these can help, they don’t have the same force as continuing to work. This makes intuitive sense—cutting investment fees is great where possible but is dwarfed by lifetime earnings and social-security benefits.

Very little, it seems, can match the benefits of delayed gratification.