> Long terms lead to more conservative hiring, leading to less placements overall.
I would argue that less higher-quality placement is a Good Thing.
> From the student side, ...
As a student, your goal is to start building your value and getting proportionally compensated as quickly as possible. Since most value as an engineer is gained well after the first few months of ramp-up you experience at every new job, you end up getting very minimal value out of a bunch of 4 month terms. Not to mention you won't be getting paid nearly as well as someone who is not labelled as a "student".
18 months is marginally better because you're actually growing your skillset, integrate as an important and dependable part of the team, and gain valuable engineering experience. On the other hand, you're still under-compensated and it's trickier to switch jobs half way through if you realize that you've made a mistake.
Finally, my argument is that the most optimal path is skip coop/PEY and find a relevant job right out of school. Even if it's not the best job in the world. You'll be far better off sticking around for 12-18 months, quitting, and getting a second real job, than to do 18 months of PEY and getting your first real job. Both in terms of real valuable experience and compensation.
When big companies look at your "industry experience", the counter doesn't start until you're no longer a student. I've been building and selling software since I was 14, but none of that went into account with HR of big companies when they were deriving the offer for my salary or position. Negotiation and positioning, on the other hand, did make a big difference. But that's another story.
The best value for coop/PEY/internships is to students who lack confidence or experience in finding a job right out of school, or they simply have no idea what they want to do when they grow up or want to drag out their student status for as long as they can. These are all very valid reasons.