The figure is just a little tricky to interpret. The horizontal axis shows birth year, and more specifically, it refers to those born in 1933, 1943, 1953, and so on up to 1993. The vertical axis show median job tenure with current employer. And the lines show a breakdown by age group. Thus, the top orange line shows that workers who were born in 1933 had a median tenure at their current job of 13 years in their 50s (which would have happened in the 1980s), but workers born in 1963 had a median job tenure of less than 9 years when they reached their 50s (which would have happened just a couple of years ago). The bottom yellow line shows that workers who were born in 1953 had a median tenure at their job of four years when they were in their 20s (that is, during the 1970s), but those born in 1983 had a median job tenure of only two years (in the 2000s).
Thus, the lines from top to bottom show that older workers consistently had more job tenure than younger workers, as one would expect. The downward slope of the lines means that those born more recently have less job tenure--if you look at the same age group.
The figure is really just one more way of confirming what most people already know: a connection between a worker and an employer doesn't last as long as it used to. Indeed, I've recently offered some evidence that in the last decade or so "All the Job Growth is in `Alternative' Jobs" (April 11, 2016), which are jobs that are temporary or on-call or "gig economy" jobs that aren't premised on an ongoing relationship between employer and employee.
For the workers, this pattern means that you need to plan your work life with a little less attachment to your current employer, and one eye looking ahead to the next job. This means continual networking about jobs. It also means that employers are less likely to offer training, because they aren't expecting to be connected with employees for the long haul, so you need to be sure that you are always updating your own training and human capital.
For public policy, falling job tenure and the rise of alternative jobs point to the importance of having benefits like retirement accounts and health care be easily portable across jobs. It also suggests that as more people are switching jobs more often, the US should consider less emphasis on "passive" labor market policies like paying unemployment benefits and putting more emphasis on "active" labor market policies for job search, retraining, subsidized employment programs, and even in some cases direct job creation to bridge the period between other jobs.
