(Population Today, May/June 2002) The Enron scandal has highlighted the issue of retirement security achieved through savings and investments inside and outside of pension plans. But the scandal aside, workers’ ability to supplement Social Security income enough to maintain their standard of living in retirement depends on the number of years they work (how much they can save) and the number they spend in retirement (how far their savings will go).

Work Lives

Trends in years of work since 1950 have been quite different for men and women (see table). There was a drop in men’s average work life from about 39 years in 1950 to 36 years in the early 1990s, followed by a reversal to about 37 years by the late 1990s. Women’s average work life, however, rose steadily and rapidly throughout the period, from about 14 years in 1950 to nearly 32 years by the late 1990s. Thus, whereas women’s work life was little more than one-third that of men in 1950, by the late 1990s it had grown to 86 percent of men’s.

Years of Work Life Expected at Age 20, by Sex and Education, Selected Dates(a)

Total Less Than High School High School and Some College Degree College Degree
1950 39.2(b)
1970 37.3
1979-1980(c) 36.8 32.6 37.6 39.5
1992-1993 36.0 29.5 36.4 39.2
1997-1998 36.9 31.9 36.5 39.7(d)
1950 14.0(b)
1970 21.3
1979-1980(c) 27.2 20.2 27.7 31.6
1992-1993 30.0 19.5 30.8 34.8
1997-1998 31.6 23.1 31.9 34.9(d)

Note: Data by education are not available prior to 1979-1980.
(a) All data except those for 1950 were calculated by the multistate method. The 1950 data were obtained by the conventional method. Tables of working life are an extension of life tables, taking into account labor force status as well as mortality. The conventional method uses age-specific labor force participation rates at a single point in time, and it assumes (unrealistically) that men enter and leave the labor force only once during their lives and that women do so only a little more often, usually when they marry or have children. The more realistic multistate (also called the “increment-decrement”) method is based on the extent to which adults at each age move into or out of the labor force or die during a specified period, such as a year.
(b) Adjusted to a level comparable to the multistate method by using a ratio of the multistate to the conventional estimates available for 1970. The ratio for men was 0.9467 and for women 0.9682.
(c) The education categories for the 1979-1980 data are: less than high school, high school to 14 years, and 15 years or more of schooling.
(d) At age 21.

Sources: For 1950, 1970, and 1979-1980: U.S. Bureau of Labor Statistics, “Worklife Estimates: Race and Education,” Bulletin 2254, February 1986. For 1992-1993 and 1997-1998: Two articles by James Ciecka, Thomas Donley, and Jerry Goldman, “A Markov Process Model of Worklife Expectancies Based on Labor Market Activity in 1992-1993” (“1997-1998”), Journal of Legal Economics, Winter 1995 and Winter 1999-2000.

Moreover, higher levels of education were associated with longer work life. The differences shown in the table between the high and low levels of education varied by seven to 10 years among men and by 11 to 15 years among women. The most important feature of these data is that people who did not finish high school have much shorter work lives than those who did finish. Since little education leads to low earnings, nonfinishers are doubly disadvantaged in their efforts to accumulate an adequate retirement nest egg. Their low earnings make it harder to save, and they have fewer years in which to do so. Although the percentage of people ages 25 and over with at least a high school diploma has increased in recent decades, workers with limited education still find it difficult to save for retirement. Poorly educated women are particularly disadvantaged.


How long retirees will need their nest eggs depends on when they leave the work force. Using estimates of the median age at exit from the labor force among workers 50 years old or older and then calculating from life tables their average remaining life expectancy at that median age provides estimates of the median duration of retirement. Between the early 1950s and the late 1990s, this duration rose from 12 years to 18 years among men (a 50 percent gain) and from 13.6 years to 22 years among women (an increase of 62 percent). (For more on trends in age at retirement, see Population Today, April 2002.)

It seems reasonable, therefore, to conclude that poorly educated workers, who have comparatively short work lives, experience a longer retirement than better-educated workers, who have longer work lives. This means that the poorly educated workers are burdened with yet another disadvantage: the need to make their retirement resources last for a greater number of years.

Challenging Combination

The ratio of years worked to years spent in retirement encapsulates the task that workers confront in accumulating adequate retirement resources (see figure). Men’s ratio fell greatly from about three and one-third years of work life for each year of retirement in 1950 to barely more than two years in the late 1990s. Women’s ratio, on the other hand, rose from about one year of work life per year of retirement in 1950 to nearly a year and a half in the late 1990s as the very large gain (about 125 percent) in their work lives outpaced the large increase (62 percent) in the length of their retirements.

Ratio of Years Worked to Years Spent in Retirement

Sources: Murray Gendell, “Retirement Age Declines Again in 1990s,” Monthly Labor Review (October 2001); and sources listed under table above.

Thus, women workers now have on average about a year and a half of earnings for each year of retirement, while men have about two years. They thus face a considerable challenge to use these years of earnings effectively to prepare for periods of retirement that are already long and may get longer. Many of these men and women live together and can share their resources, but many do not. Much depends on future trends in the length of work life and of retirement, trends that are not readily discernible.

Murray Gendell is senior research associate at Georgetown University’s Center for Population Research.

For More Information

For more on trends in age at retirement, see “Boomers’ Retirement Wave Likely to Begin in Just 6 Years,” Population Today, April 2002.