I wonder what the upsides are of hiring assistant professors instead of more experienced professors.
I am mostly interested in US universities and computer science departments.
I'm sure others can come up with a number of additional reasons.
Assistant professors are not inferior versions of full professors. Aside from years of experience, there will usually be many differences between two candidates. Young does not mean not as good (it will when I celebrate my 60-th birthday later this year, though). The assistant professor may hold an advantage in a specific field or in general.
A department may be intentionally focusing on a very specialized subject field that the assistant professor brings. That is to say, an assistant professor may be more proficient than a full professor in some departmental subject matter.
For a more general illustration, I believe I was better at one aspect of teaching when I was a young assistant professor than I am now. I was better at helping students break through brick walls (such as clearly understanding the concept and point of calculus) when I personally could "remember not knowing". It is easier to explain how to go from not knowing to knowing when your own breakthrough was just a few years earlier instead of many decades earlier.
In a system like the US, a "more experienced professor" typically has tenure already. Someone who has tenure won't want to give it up. Therefore, if you want to recruit them, you have to offer them tenure as well, either from the moment they walk in the door or very shortly thereafter. Otherwise they are not going to take the job you are offering.
This creates a significant risk for your institution. The decision about whether to hire this person and grant them tenure has to be made based on limited information: their CV and publications, some letters of recommendation, and a one- or two-day in-person interview (during which the candidate is of course on their best behavior). These materials may not be representative of how the candidate will actually perform in the job. Maybe their teaching turns out to be mediocre, or not a good fit for your student population. Maybe they shirk all administrative duties. Maybe some aspect of your institution makes them less productive than previously. Maybe they have catastrophic personality conflicts with their new colleagues. Too bad for you! You gave them tenure, and now you're stuck with them until they retire or leave voluntarily.
On the other hand, if you hire an assistant professor, you have five or six years to observe their actual job performance close up, before deciding whether to tenure them. You don't have to extrapolate from what they have done at other institutions - you know exactly how they work here. You know how effectively they teach your students. You know how well they get along with people in your department. If any of this turns out to be unacceptable, or even if it's okay but you think you could find someone better, you can fire them (or deny them tenure). So your risk of getting stuck with an ineffective or toxic faculty member is quite a bit lower.
(Note: "You" in this answer could refer to a dean, department chair, or to the faculty of a department or institution collectively. Basically, whoever has the responsibility of making hiring, retention, and tenure decisions.)
In addition to the other answers already given, there are simply not enough experienced professors on the market to fill the demand for professors.
Let us consider the system of academia from a global perspective.
There are two general pools from which professors can be hired:
Thus, the only way to increase the net number of professors enough without hiring a lot of assistant professors is if there are a large number of people with associate-professor-level qualifications, despite not having ever been assistant professors.
Given that many people who get doctorates have no desire to become professors, it seems quite implausible to me that there would be such a large pool of people that can move outside of academia gain massive amounts of experience, and them move back in as higher-ranked professors. Moreover, such people would likely be higher age (since their primary job would generally not be gaining professor-type experience and qualifications), which would further increase the fraction retiring each year, thereby further intensifying the population pressure.
Thus, from the point of view of pure population analysis, there is an extremely high pressure for academia to hire at the assistant professor level. Individual institutions may buck this trend, but overall it is likely to hold quite strongly.
Note that the numbers in this analysis are US-only, but if we extend to international the impulse to hire assistant professors will likely be even higher, as academia is expanding much more rapidly in the developing world.
In short: perhaps the biggest benefit of hiring assistant professors is that they are available to be hired, whereas the pigeonhole principle implies that for most institutions it will simply be impossible to fulfill their faculty needs without hiring many assistant professors.
In addition to Corvus's answer, part of a researcher's motivation to publish his/her research is to gain departmental promotions (associate professor, professor). By starting at a lower ranked position, the faculty member will more likely publish more papers under the university's affiliation during his academic career (on the premise that most faculty will have reached the professor rank by the time they retire). More publications under the university's affiliation will directly affect the university's ranking and external grants and funding.
This of course depends on the department's promotion policy but most will weigh publications as the highest criteria.
I think there are a number of excellent answers here and some interesting discussion.
Let me add an additional point for the "middle tier" schools or department. If we hire an established professor, we have demonstrated that this person:
In short, when a mid-tier school hires at the senior level, word gets out and other schools may try to hire instead. This might happen at the same time we try to move them, or at some point later. After all, if someone is willing to move once, they are likely to move again for the right offer.
If you're at the top of the reputation ladder, you can easily find mid-career or senior talent at other schools and lure them to join your department. There's incentive to join your department because it's "better" for some definition of the word. You also likely have fairly deep pockets for start-up budgets and salary.
Consider instead the case of a good-but-not top department. If we find a promising mid-career or senior candidate, we have to compete not only with other mid-tier programs but also with the possibility that a rising star may be poached by a top department.
In short, when you hire assistant professors, in addition to the answers elsewhere, they start with some loyalty to the department and the school. (You gave them a chance!) If you try to hire an established professor, you've already proven that they're willing to leave a school for a better offer.