Short-form Answers: 1) We need to be careful not to perpetuate the idea that UX design represents a phase that can be researched, completed, and finished. UX design is a process -- not a phase. 2) User research can be done in any phase of product design and development, but the value of user research in some phases is far more important than in others.
In my experience, I like to think of UX design as an integral part of all phases of a Software Product's life cycle. I don't view UX design as a single "phase"; nor do I view UX design as existing without the context of the various phases of software development. For example, there are UX considerations in forming a product vision; there are UX considerations in identifying target user groups; there are UX considerations in establishing requirements; there are UX considerations in developing an internal data model etc...
The methods used to fulfill the UX design goals of any given phase of product development are specific to that phase -- and each requires its own kind of work.
Certain phases, such as "requirements gathering", absolutely and significantly benefit from user research. Phases such as "validating individual screens are parsed as intended" aren't even possible without involving other people.
On the other hand, in phases such as "forming a product vision", the biggest priority has to do with company strategy: "What angle of this market do we want to go after?" That said, hopefully the people making these strategic decisions have done some of their own research to back up their work.
Ultimately, doing user research involves cost and provides benefit, and the benefits of user research accrue differently for different phases of a project. And the reality is that sometimes we need to do triage, and pick and choose where and when we apply our resources