This paper uses longitudinal data to estimate cohort changes in the ea
rnings trajectories of young adult males. Levels of earnings are unifo
rmly lower for male workers turning 21 between 1980 and 1991 than in 1
970-1979, although rates of earnings growth are roughly comparable. Am
ong males turning 21 before 1980, six in 10 (60%) of all men and seven
in 10 (71%) college-educated men attained earnings levels by age 30 t
hat were at least twice the poverty level. Corresponding fractions for
workers turning 21 between 1980 and 1991 were considerably lower (42%
and 56%). Recent cohorts from all demographic subgroups appeared to h
ave more difficulty than older cohorts iii attaining middle-class earn
ings.