Inside Higher Ed has an opinion essay by UToronto’s Mark Kingwell arguing against the institution of tenure, or rather arguing that it should be revisited. Some of the comments are as interesting as the article itself, but it’s worth a read.
As an aside, I disagree with Kingwell, but I don’t have a fully articulated critique (yet).
An Ergodic Walk 
The last (current) comment by “Paul” is the most accurate, I think; regardless of any arguments for or against tenure and how well it works, it’s an economic proposition. Tenure is valuable. I took and am taking a lower salary than my peers in industry in favor of several non-monetary benefits, one of the most important of which is the opportunity to earn tenure. If tenure was not available, salaries would have to be higher to compensate (if universities were to keep the same caliber of professors). For good or bad, universities have chosen to award lifetime employment in lieu of a higher salary.
Yeah, I found his comment the most persuasive as well — but the analogy seems to work best in engineering, where there are lucrative industry jobs out there with less security.
Is the same true in history, e.g.? I doubt it…
Well, history PhDs don’t have the same breadth of lucrative industry options that we do in EECS, but the calculus remains the same: tenure is still valuable and if it’s removed without compensation, the job is less attractive. For history, the result would perhaps be fewer (or less qualified) grad students entering history PhD programs, instead choosing to go to law school or other non-academic routes.
Catching up with your blog, after a while. I loved your math/stat interludes… Regarding tenure it is a very difficult question indeed. One thing is for sure, not having tenure will mean universities have to spend more to hire and keep the best talent. Not only because they might prefer industry research, but also other universities now can make much easier offers… How much more? I am guessing at least as much as the difference between associate and full professor.
It also becomes unclear the advantages of working for 10 years, on funding you bring to your university. A simple discounted cash flow analysis values a prof position values it at $2 million for a $100K salary at 5% return. Typical NSF career awards are around 400K. So you can see that in some sense, you already do a enormous down payment on your debt to the university… In fact, most EE profs will be bringing in more than $2 million in their career. So from the institutional point of view, you are zero net, and for yourself, you got labor and support. The major condition is that a risk free 100K are guaranteed for you for life, after tenure.
Without this condition, why not have your own private research institutions with smaller overheads? Which leads to another question: riskier research will not have its risk mitigated by bigger money makers on campus… At the end of the day, somewhat commercially interesting research can survive, but foundational research on social sciences, for example, will suffer a lot…
My guess is tenure is a mechanism to make us all accept a certain form of socialism: I guarantee you 100K for life, for a certain initial level of output, and in return, you share the research grant money you bring (which is usually taxed at 50% by the institution)… Thus, a literature prof. is certainly benefiting some from your grant…
If there are better systems? Maybe. But I would certainly skip teaching altogether and departmental meetings, if there was no sense of stability and community in an academic environment…