Twenty years and one month ago, on 1 January 1994, it became illegal for universities to force tenured professors to retire. Last August, while writing about a study of that change's effects, I repeated the widely held idea of a generational conflict on campus. Faculty members in their 70s and even 80s who decline to retire, this argument goes, block the way to tenure-track jobs for the coming generation of young academics. I have since learned, however, that in science departments at research universities (as well as institutions aspiring to be research universities) the situation is more complicated than that simple analysis suggests.

At first glance, it makes intuitive sense that the salaries devoted to the growing ranks of expensive (and sometimes no longer productive) senior professors could support larger numbers of cheaper and more energetic junior faculty. The number of professors eligible for retirement increased twofold from 2000 to 2011, according to Bureau of Labor Statistics figures cited by The Chronicle of Higher Education. On some campuses, a third of tenured professors have reached age 60. That's a lot of potential job opportunities.

Some older faculty members continue to work because their pension funds took big hits in the financial collapse of 2008. Others, it seems, simply like their jobs too much to leave: In a 2009 survey, 66% of University of California, Berkeley, faculty agreed or strongly agreed with the statement, "My main satisfaction in life comes from my work." In the U.S. workforce at large, only about 30% gave those answers.

But at the universities that expect their science faculty to compete successfully for research funding, it's typically better for the department if those older faculty members stay on. If all of their retirement-eligible tenured scientists were to take their pensions and decamp to rocking chairs, lots of these universities would have great difficulty bringing on assistant professors to replace them. Like much else bedeviling the career prospects of aspiring academics, the reason lies in the contorted structure of research funding.

The competition for likely winners

Most research universities would like to bring new talent into their science departments, but new scientists are expensive. Yes, they're paid less than their senior colleagues, but the most important expense for universities isn't salaries; it's startup costs. The packages that research institutions must supply to prepare junior faculty to compete for—and win—serious grant funding now run "upwards of half a million dollars" per new assistant professor, explains Linda Abriola, dean of the engineering school at Tufts University in Medford, Massachusetts, in an interview with Science Careers. "I believe that that's true in all the sciences and engineering with the exception of those that are just computational," she continues. "For a biomedical engineer [or similar life-sciences investigator] who is doing laboratory-based work, this number could be over a million dollars."

Given the current brutal competition for grants, the math seems inexorable. "When we hire tenure track faculty, the expectation is that you give them student support, summer salary, reduced teaching loads, money for equipment, additional space," Abriola says. "We commit … to four student years, for example. [At] $50,000 per student per year, that's $200,000 right there." That's before buying any equipment.

Senior science faculty members pull down bigger paychecks than new assistant professors, but they generally end up costing the university less. The older people, Abriola says, "are bringing in grant money … to support students" and other costs of their labs, often including at least part of their own salaries, not to mention overhead payments for the university.

"Twenty five years ago the startup packages were just not as large," Abriola continues. Also since then, the odds of a young Ph.D. landing a tenure-track job have dropped substantially. "It's astonishing to me the amount of resources … being put into recruitment of junior faculty" these days, she says. What has changed? "It is harder to get research funding now, so these junior folks need resources to get their initial data and get their research programs started so they can be more competitive" in the contest to win the research grants that justify the university's large initial investment.

Paradoxically, in a labor market glutted with young Ph.D. scientists, sharper competition for research grants has sharpened the competition among universities to sign the likeliest winners. "There's a certain group of individuals … who are very competitive in terms of their credentials," Abriola continues. Because these people appear to have the best chance of bringing home funding, "the types of startup packages that [they] are getting are very large. … If we don't offer these kinds of packages, we are not competitive" for the most sought-after candidates. Most in demand, she explains, are those few elite young scientists who have already attracted highly prestigious support from such sources as the National Institutes of Health's K99/R00 "kangaroo" grants or the Howard Hughes Medical Institute's Investigator Program. "They can go anywhere they want," Abriola says.

So, to bring on inexperienced investigators to replace retiring ones, research universities have to commit large amounts of their own money. "Twenty percent of our faculty [members at Tufts] are eligible for retirement. … If all of those folks were to decide to retire, it would take us years to rebuild."

Given these economic realities, some departments choose to find new faculty members by luring established investigators from other institutions. Though this can also cost money, these older newcomers bring their labs and the grants that support them, and they have already proven their ability to win new grants. Alternatively, universities needing to fill retirement-caused holes in the instructional schedule may find it "much more economical to replace" a senior person "with a part-time lecturer or even a full-time lecturer" than with a tenure-track assistant professor. "There are no startup costs associated with it," Abriola says, "and the salaries of lecturers are typically lower, obviously, than the salaries of full professors."

Age and productivity

"A lot of senior faculty are extremely productive, which means we'd like them to stay as long as they are productive," Abriola says. That's a big change from the days of mandatory retirement. In 1987, a chemist named John Fenn turned 70 and ran up against Yale University's then-strict retirement rule. Involuntarily moved to emeritus status, he retained an office, but his lab space shrunk. This "made me angry because [my research group was] right in the middle of things, and so I fought tooth and nail, but the bureaucracy was pretty adamant. I got tired of fighting," Fenn said in a 2009 interview.

When the law changed in 1994, Fenn relocated his lab, becoming a professor at Virginia Commonwealth University (VCU) in Richmond. That was his academic affiliation when he shared the 2002 Nobel Prize in chemistry, for research in mass spectrometry—work published after his forced retirement from Yale. He continued working "nearly every day" well into his 90s, according to a VCU colleague.

Though Fenn, who died in 2010, is an exceptional case, highly productive scientists at or beyond retirement age are hardly rare. Nor are research universities like Tufts representative of the thousands of nonresearch institutions that employ science professors. At colleges where faculty give their main attention to teaching rather than research, and therefore stay out of high-level competition for grants, the startup funding departments provide doesn't cost as much. In those circumstances, senior faculty who linger can, indeed, create what The Chronicle of Higher Education calls a "faculty bottleneck" that blocks job opportunities for younger professors. At research universities, though, even if retirement-age faculty were to retire en masse, their departments would have difficulty replacing them.

Beryl Lieff Benderly writes from Washington, D.C.

10.1126/science.caredit.a1400033