Journal submission guidelines in my field typically require the submitting author to provide the editor with a list of potential referees. The journal guidelines ask to suggest people with whom we have no collaboration, no joint publications and no conflict of interest.
Given how competitive academic research is, I sometimes wonder how conservative one should be in choosing potential referees. After a long career, if I am to exclude all people whom I have collaborated with over time, there will be few experts left qualified to judge the work! On the other hand, someone with whom I am just starting a project with, and has nothing published with me so far, would not be seen as a violation by the editor, yet would not be rally fair (if we work together, he has a vested interest in my success).
So, I am asking here what guidelines or principles can be proposed in this gray area.