The incentives vary strongly by location, although I think the usual driver is monetary. For instance, in Germany, it is difficult to be employed as a PhD student at a single university for more than six years (following the completion of the master's degree). Moreover, because the recruiting of PhD students is often done "from within," if a professor becomes known for not graduating students, it can hurt long-term recruiting. (Also, DFG funding typically can be extended on a single project for five years.)
For US professors, there may be an expectation of graduating a certain number of students as part of the requirements for being granted tenure. However, I think the bigger incentive is that beyond a certain point it becomes too difficult to keep finding funding for a long-term student.