Against this array of forces, longer survival and reduced morbidity push in the opposite
direction, tending to raise labor force participation.
For example, I estimate that improved survival since 1990 increased working lives by
about 2 years in China, holding age-specific labor force participation rates constant.
Consider the counterfactual XLFP calculated using 1990 sex-specific survival rates to
each age, but holding age- and sex-specific labor force participation rates constant at their
2010 values. If rural men faced risk of death as in 1990, their 2010 work-life would have
been 2.4 years shorter because of premature mortality, but work-life would represent a
higher share of life expectancy at birth (by a little over 1 percent). For urban men with
1990 survival, working lives would have been 1.5 years shorter. The corresponding
numbers for rural and urban women are 2.3 and 1.2 years, respectively. Thus, improving
survival contributed nontrivially to expected labor force participation increases in China,
ceteris paribus.
A critical question for the future will be how age-specific labor force participation rates
evolve. Although our analyses are mostly “just” an accounting exercise, the importance
of behavioral change can be highlighted by such decompositions of the changes in
expected labor force participation into changes in survival and age-specific labor force
participation. For example, consider an analyst of the US economy in 1900 who naively
assumes constant age-specific labor force participation rates into the future. Projections
of XLFP would have been widely off the mark, even if correctly foreseeing increased
survival. In particular, such an accounting approach would have underestimated the true
increase in female XLFP by 90 percent, and overestimated the increase in male XLFP by
80 percent, because age- and sex-specific labor force participation rates changed
significantly over the 20th century (Eggleston and Fuchs 2012).
Further evidence regarding the evolution of work-lives comes from estimates of
consumption and income by age from National Transfer Accounts (www.ntaccounts.org),
a project that Ronald Lee has pioneered and covers in greater detail in his contribution to
this volume. Chen and Lee (2014) show that in 1995, the “cutting ages”—ages where the
income curve cuts the consumption curve—were 20 and 60, for a working life of 40 years.
By 2009 (the latest NTA available for China), the cutting ages were 21 and 56, for a 35-
year work-life—although more Chinese were surviving into the working years and
beyond.7 In India, the corresponding cutting ages were 27 and 59, for a “work-life” of 32
years (Chen and Lee 2014)—despite India’s younger demographic, more rural populace,
and lower investments in education, especially for girls. For comparison, in the US in
2003, the cutting ages were 26 and 59, representing a 33-year “work-life” of income
exceeding consumption (Chen and Lee 2014).