Research environment is over-rated

For assessing grant applications in Australia, and some other countries, one criteria is "research environment". This means different things to different people. Unfortunately, I too often see both applicants and assessors/referees using this criteria in an unhelpful and/or meaningless way.

When does the environment of the proposed research project matter?
Here are a few ways, listed in order of decreasing importance.

Access to crucial equipment, materials, and infrastructure.
For example, if the project involves femtosecond laser spectroscopy, then there is little point if the researchers do not have access to the relevant lasers, probably at their own institution.
Similarly, neutron scattering requires relevant beam time on a user facility. Experimental studies of strongly correlated electron materials require access to high quality single crystals of the relevant materials. Large scale computational chemistry requires access to the relevant supercomputing facilities.

Access to intellectual resources.
Colleagues with relevant technical expertise may enhance the project. Also, a theory (experimental) project can be enhanced by the local presence of an experimental (theory) group that is actively interested in similar problems and systems.

A lively and interactive department with a history of fostering new collaborations.

What are debatable measures of research environment?
The overall "ranking" of the institution that will host the grant. Previously, I posted about a study that investigated whether moving to a more highly ranked institution improved research quality [as measured by citations].
The "ranking" of the host department in some silly national "research quality assessment" exercise.
The geographic proximity of "high profile" research groups or "big fancy shiny buildings" with $M budgets in vaguely related research areas.

What do you think?
How does research environment help research quality?