But before I start, two caveats:
- the outcome of a job candidate is not determined entirely by the placement efforts of individuals and departments--there are both the individual candidate's qualities and performance in play and the politics of the hiring department.
- there are departments and individual advisers who suck at this.
OK, the department logic is pretty simple: the reputation of a department hangs on a few things, with the placement of their grad students being a key ingredient. If departments do not place their students, the word does get out to a degree, making it harder (although not impossible) to attract good graduate students. It might also make it harder to hire faculty since people tend to prefer to be at places with upward trajectories. It is also something that Deans care about, as they prefer to have departments that do well in the rankings than departments that do poorly. They can direct resources to and away from departments, so their concern about rankings is important.
While I have often griped here about reputation and rankings--that they matter less than people think--it is pretty clear that people do care. There are other ingredients, but getting your students placed at all and especially in more prestigious places is a way for a department to change its reputation.
The individual level requires me to borrow a distinction that I dissed long ago: instrumental and affective motivations. It is both good for one's career to place one's students and good for one's ego/guilt/whatever.
Having successful students--those who gets placed (narrow definition for the moment) is good for one's career. Really. How so? First, it makes one more marketable as departments seeking to improve their ranking might want to hire profs with records of supervising successful grad students. Second, grant giving agencies often ask about one's record of working with grad students (just submitted such an application). Third, departments sometimes have incentives for successful supervision. At my last job, supervising Phd students produced points that would eventually accrue to a level that would allow one to get a course reduction. Fourth, places that have merit increases may include finished students as one part of the calculation. Yes, finishing is not placing, but they are not unrelated.
Perhaps more important are the "affective incentives." That agreeing to supervise someone is serious business, that it represents a commitment to not just read a few chapters but to mentor someone. The idea is that the grad student is an apprentice that one is supposed to get not just through grad school but through the early stages of their career. So, one should should help place one's students because it is part of the responsibility, part of the unstated contract. Then there is the reality that when one works with a student for three or more years (from dissertation proposal to dissertation defense), you tend to develop a connection. That you want them to do well because they are not just random students but people you care about. This is what distinguishes PhD students from any other--the length of the relationship.
Again, to be clear, there is only so much an adviser can do. They can provide feedback on the dissertation and on the articles that the student writes along the way, advice about the CV and the cover letters. Of course, a key task is to write good letters of recommendation (I have seen a big name or two write letters that are less than three paragraphs--not good. What counts as a good letter is a subject for another day). And, yes, advisers can contact search committees, but I am not sure how much that matters. I do not contact people I do not know, but I do email or call people I do know to plug my students. It might help them get from the pile of 100-200 to a long short list of 25 ... or not.
But other factors matter a great deal--does the candidate fit what the department needs? Does the candidate have a strong record? Do competitors have better records/better fit? And then politics comes into play. Once that stuff comes into play, then a short list is decided. If one's student gets an invite, then it is entirely out of the hands of the adviser. The key variables then become how the candidate does in the job talk, how well they interact with the various people they meet (some schools will have a committee meet with a candidate and seriously drill them for an hour or two), how the candidate fits into the political dynamics of the place, and, of course, how well the competitors do at these various things.
Of course, students will worry about what their advisers are doing since this is something that is out of their control and has much mystery attached. I used to whine that my adviser was not doing enough, but my early struggles probably had much more to do with going out on the market during a recession, my lack of publications, that my stuff was not as hot as other stuff (ethnic conflict fared poorly when compared to environmental politics in the early to mid 90s), and that incomplete was my training.
Anyhow, I have taken pride in that nearly every PhD student I have supervised was able to get tenure track position. But much of that pride is not really because of anything I did--but my ego was invested in what my students were doing. That their success became my success. So, when someone says there are no incentives, I re-flummox.