One would need to examine that nature of the referral. For example, if colleague-1 recognized that colleague-2 had "more expertise" in a particular area and could offer "better advice" than colleague-1 then this would be a consideration of referring the individual to the colleague after sharing with the individual why the referral is being made by colleague-1 to colleague-2. Negative examples of "unfair referrals" can also be developed.
I believe that an individual has an obligation to refer an individual to a "more qualified" individual that the individual patron-1 if two conditions hold: (1) patron-1 knows that patron-2 is better (through more training, more experience, etc.) than patron-2, but not necessarily refer to the best patron overall, rather referral is to the better patron-2. This would fit into most existing regulatory frameworks in many countries.
wow! wow! wow! This is a really crucial point nowadays. I suspect that is a very common practice everywhere. A lots of journals are asking you reviewers as a mandatory, this aspect must be removed as a mandatory. This is one of the big problem. To be an Editor or a Reviewer it can became a job nowadays.
For editors, it is ethical to involve the best available specific experts, no matter whether they are friends or "enemies". As an editor, I appreciate reviewers suggestions by authors. However, at the end, I base my choice of inviting someone or someone else on my perception of potential reviewer's expertise.