Europe has already pushed shorter PhDs with the excuse of facilitating development in technology/industry. It's been a huge failure in mathematics for various reasons among which are: 1. Students don't have time to develop autonomy and independence, so this remains for the postdoctoral period. 2. Students are forced into publishing things that shouldn't be published (it's often a requirement to terminate a doctorate) and these are shallower and less complete than they should be.

The true motivation seems to be that it is cheaper to fund 3 years than 5. Master's degrees have been made obligatory prior to doctorates (in the US this is traditionally not the case) and so are not covered by doctoral funding - rather they are paid for by students in many cases.

It may very well have sense to have a shorter track aimed at industry, but it shouldn't be called a doctorate and all doctorates shouldn't be shunted into this setup.