Editorial board members generally select reviewers, searching:
Their own contacts for relevant potential reviewers including talks heard at conferences and meetings, and journal articles read:
databases such as those provided by the journal management system/publishing house and the
web for researchers with relevant experience.
In my experience, the databases turned up a mixed bag of suggestions, mostly comprised of a stream of irrelevant and/or incompetent potential reviewers, so many of these suggestions ought to be eliminated. I found that searching the web, then checking the publications lists and research interests of potential reviewers was more fruitful than wasting time with the databases because so many entries were associated with, for instance, a one-off conference paper rather than the depth of experience one expects to find in a reviewer. Given that researchers identified through searches should be checked out to make sure that they have the relevant experience and background to provide a competent review of the manuscript under consideration, a certain level of expertise is required to choose reviewers. This can be problematic as many journals have a policy to use only one of the reviewers suggested by authors, so the longer the list of suitable reviewers provided by authors, the more potentially suitable reviewers are excluded. Having said all this, understanding the process will probably clarify your observation: Editorial board members are asked to provide a list of at least five potential reviewers, if these people decline to review, and the board member does not get back with a second lengthy set of suggestions (and it takes time to put together such a list when selecting with care), the journal staff will prefer to select reviewers than wait for the board member. This is because one means of assessing the editorial staff's performance is how long it takes them to obtain decisions on the reviews of the papers that they are overseeing. If you are systematically finding that one of the reviews returned is irrelevant or demonstrates incompetence in the field, then it could be the case that a member of the journal's staff selected the reviewer, not realising that the result would be an incompetent review. Depending on the journal and publishing house, it is quite possible that the publishing house is employing an editor who does not hold any qualification in the field covered by the journal. Making the choice of a competent reviewer under such circumstances must be problematic for them.
So how could this situation be improved?
The databases if people approached to review a paper that falls outside their area of expertise were to return a standard letter that they could keep for the purpose explaining their area of interest, providing keywords and clarifying the papers that they would (and would not) be prepared to review. This information would then be added to the database.
Publishing houses could stop rewarding staff for taking shortcuts.
Authors of papers might be able to improve their suggestions for reviewers and provide useful assistance for journal staff.
How could you assist journal staff:
Provide a list of a few suitable reviewers, preferably including some less known researchers (for instance post docs) and without making the list exhaustive! You should exclude colleagues with whom you have published recently (however if there is a special reason to invite such a person – like the person is indisputably the best in the field, you could argue for this in a cover letter);
Make sure that your reference list contains relevant references as researchers or colleagues of researchers whose names figure in the list might be a useful source of potential reviewers and
Consider the keywords that could be used in the selection of relevant reviewers – does this list overlap with the keywords for your paper? If not, why not? Would researchers in a complementary field have the expertise to review your work and be able to shed light on the interpretation?
Where a review is completely incompetent, respond to each point raised carefully, but make a note to the editor that you consider the review to have been inappropriate and unhelpful for the quality of your work.
The editors are probably selecting the reviewers from the list of editorial board, or they have a separate list of past reviewers available. The choice of reviewers is not always commendable.