Steven Bainbridge, in the course of congratulating Steven Taylor on his promotion, makes the following observation:
When I was up for tenure (a nerve-wracking time, even worse than sweating out the bar exam), a senior colleague told me that getting tenure didn’t change anything in your life except that you stopped thinking about tenure. I didn’t believe him, but it turned out to be true. If you’re internalized the norms of teaching and scholarship, you don’t change what you do. You just keep teaching and writing.
This seems like an odd analysis; the grant of tenure* doesn’t remove the incentive to publish, teach, and perform service at a high level (in the various department and college-prescribed ratios); while it is true you can no longer be fired for failing to do those things as proficiently, most associate-level professors at least aspire to promotion to full professor and the prestige and monetary rewards associated with that rank, which requires a similar level of effort (as between assistant and associate) to attain. Thus, we would rationally expect that professors would be more likely to slack off after promotion to full professor, rather than after achieving tenure.
This would also tend to coincide with the observation that scholars tend to produce their greatest contributions as researchers in their first 10–15 years after the terminal degree—a timetable that more accurately reflects the typical schedule for promotion to full professor than to associate. Lengthening the promotion-and-tenure schedule might lead to greater per capita contributions. Plus, if Prof. Bainbridge’s suggestion that the tenure process causes internalization of norms is accurate, a longer process might be more effective in fostering those norms.
Of course, there are advantages to the existing structure: most notably as it reduces the incentive for more senior scholars to contribute research to the more prestigious journals that junior scholars need publication in for promotion.
A good introduction to academic ranks and the tenure process is here.