where for men than for women. The average probability of moving up is also
highest among younger workers and thereafter declines with age for both men
and women. The average probability of moving down is highest for entry and
older workers and lowest for prime age workers, again for both sexes. The
average net probability of moving up for men is highest for younger workers
and thereafter declines with age, but for women is highest among prime age
workers. The average net probability is everywhere greater for men than for
women across all age groups, but the differential is greatest among younger
and prime age workers (8.1 - 10.2 points) and lower among entry and older
workers (6.0 - 6.1 points). Again, this is consistent with men having generally
higher and more concave career earnings trajectories than women.
Net responsiveness of the summary mobility measures by age group is
examined in Beach and Finnie (2001). For men, the sensitivity of earnings
mobility measures to changes in unemployment decreases monotonically with
age. The average probability of moving up is fairly cyclically responsive for
all ages attenuating from a marginal unemployment rate effect of -0.82 for
entry workers to -0.28 for older workers. When the unemployment rate rises,
the probability of moving up falls. The probability of moving down also
shows significant cyclical responsiveness declining from 0.62 for entry
workers to 0.27 for older workers — it rises along with the unemployment
rate. But its strength is generally less than that of the upward mobility. As a
result, the net probability of moving up shows significant cyclical variation
among all age groups of male workers moderating from -1.44 for entry
workers (-0.99 for younger workers and -0.81 for prime age workers) to
-0.55 for older workers. Between 50 and 60 per cent of the net upward
mobility cyclical effect arises from changes in the probability of moving up
across the four age groups of male workers. But perhaps most interestingly,
since the shift downward in net upward mobility decreases with age, the
earnings advances of entry and young workers will be dampened much more
than those of prime age and older workers, thus widening the gap between
relatively low earners early in their careers and higher earners later in their
careers. The result is that men’s cross-sectional age-earnings profiles swivel
down during recessions and their steepness and concavity are accentuated —
as illustrated in Figure 4. In periods of economic expansion, the reverse
occurs. For female workers, the net cyclical earnings shifts are far smaller and
show no such statistically significant pattern. For example, the unemployment
rate effects on the net probability of moving up vary from -0.19 for entry
workers to 0.17 for older workers (none of which are statistically
significant). Consequently, women’s cross-sectional age-earnings profiles do
not show any significant change in shape over the business cycle.
Cyclical Changes in Short-Run Earnings Mobility in Canada
479