The incentive system in academia is to produce people who cite their mentors in formal research. That's it. Because universities are non-profits, they go to the indirect incentives. This trumps any altruistic ideals. In their defense, most Phds in computer science are fully funded, so it's not transactional like an undergrad degree.

I've worked with several Phds with background in Stats and Computer Science. I found all to have "Good Background" - they knew a lot about their fields. Some were great at working with others, but some not so much. It wasn't a magical degree. Many regretted not getting the 5 years of work experience instead. Relative to the All But Dissertation crowd, they were a little better at Getting Things Done, versus just talking about things. (Very small sample size, so don't read too much into it)