Skip to content Skip to navigation

Working Longer and Retirement Conference

Jul 2014 to Jul 2017



Alfred P. Sloan Foundation


The list of the “class one” problems for the 21st century for the American economy usually includes funding the entitlement programs, dealing with a transition away from fossil fuels to reduce the risk of climate change, and controlling the runaway costs of health care. There is one more item that needs to be added to the list. That is adjusting economic institutions to dramatically longer lifetimes. To get a sense of how much life expectancy has changed in the past fifty years, look at just one statistic – the remaining life expectancy of 65-year old American men has increased by over 40 percent during that period.

The average retirement age in the US is 62 for women and 64 for men. The most common age of retirement, i.e. the modal age, is 62 for both genders. How much longer can a 64 year-old husband and a 62 year-old wife expect to live? Social Security data indicate that the expected time until the first death for this couple is about 16 years and the expected time until the second death is about 27.5 years. The couple or at least the survivor of the couple can “look forward to” a 27.5-year retirement with average luck. But, there is a great deal of uncertainty about the length of life. Any couple might have better than average luck. If the couple does not annuitize their wealth, they have to be financially prepared for being in the upper tail of the survivorship curve. To provide a sense of the magnitude of the uncertainty about how long retirement will last for this couple, in ten percent of 64 and 62 year old married couple retirements, the survivor will be around for another 35 years or longer. In fact, one member of the household has a five percent chance of celebrating her/his 100th birthday.

How long could this couple have worked? 40 years seems like a reasonable guess, given their retirement ages of 64 and 62 and allowing for just a little bit of time off for child rearing, unemployment or other circumstances. So, we have a couple attempting to finance at least a 30-year retirement with a 40-year career. It won’t work unless the standard of living in retirement is way below pre-retirement levels or the saving rate during the work years is so high as to be distinctly un-American. 30-year retirements with 40-year careers won’t work in the US, in Europe or in China for that matter. They won’t work with public pensions and they won’t work for DB or DC plans. They won’t work for the best designed pension plan imaginable. The attempt to do the impossible (financing 30-year retirements with 40-year careers), is an important factor in the current economic strains being felt everywhere from the countries of Europe to cities of California and the midwest.

Some would look at the demographic facts just mentioned and say that Americans need to save more. That may be true, but the more feasible adjustment is to work longer. By the middle of the century, Americans will probably need to be pursuing 50-year careers with an average retirement age in the range of 70 or 72. Not everyone will need to work longer, disability institutions will have to be more robust and labor market institutions will have to change. We also need to examine the inequality of mortality progress and any institutional adjustments we make to accommodate it. We may want to examine how Japan, where people live four years longer than Americans, have dealt with the issue. With an extra 6 to 10 years in the workforce, pension institutions would have a chance to fulfill their mission. Currently, we are on “mission impossible.”

It should be added that working longer would be good for the macro economy in the long run as well as good for financing lengthy retirements. I have in mind that the economy is like a machine that produces more output when it is fed more inputs. More labor supply translates to more output (GDP) and therefore more GDP per capita and a higher standard of living. Of course, this raises the question of aggregate labor demand and the role of employers in the whole working longer policy. While the lump of labor hypothesis (that there are only so many jobs that the economy can provide and that if older workers work longer, young people will have difficulty finding jobs) has been empirically discredited, it still is a political obstacle to working longer policies. Over the course of the three conferences, I plan to put considerable emphasis on the labor demand (i.e. employer) aspect of adjusting the balance of work and retirement in America.