Can anyone direct me to some guidance documents which could clarify when a method transfer is necessary? For example, what are the guidelines when transfering methods within the company, but from one facility (R&D) to another (QC) across town?

Anything that pertains to method 'robustness' is the type of thing that gets looked at through transfer protocols.

Basically, you have a sample set divided and run by the lab that developed the method and the lab it is getting transferred to. If the results fall within the same statistical population, you're done.

You should have a look at the related discussion "Need to revalidate the analytical methods ?" started on Jan,15 2004.
My contribution was :
If a method has been developped and properly validated in a GMP lab, there is no requirement to revalidate in an other lab; however it is a good practice that the receiving lab formally documents the method transfer.
This documentation should examine weither any discrepancy between local set-up (instrument brand, type, configuration, software type, reagents grade ...) and validation conditions is already adressed in the validation file (mainly in ruggedness and robustness).
If not or partially, the transfer documentation should include validation complements (it is also a good practice that the receiving lab keeps the validation lab aware of these complements so as they are available for a future transfer in a third lab).