I would hope that no clinician would ever make a frivolous diagnosis. But certainly, erroneous diagnoses are made and possibly some that are wrongful, rather than erroneous.
Factors that can lead to erroneous diagnosis are one's training, one's practice, current trends, and incomplete information or misinformation. One's training and/or practice may expose a clinician to some conditions, but not others, for example, making it more likely for the clinician to "see" patients through the lens of that exposure. Trends seem to skew diagnosis as well. For example, some research indicated that children could be bi-polar, and before we knew it, the incidence of juvenile bi-polar diagnoses exploded. These were not all necessarily misdiagnosis, of course, but these trends do seem to skew the diagnostic process. Similarly, diagnoses of autism and ADHD have risen almost exponentially, possibly reflecting better understanding, but also possibly reflecting selective perception. Diagnoses, because we have primitive mechanisms for viewing the brain, are largely dependent upon what the client and/or others report. Thus, any information withheld or not reported accurately can certainly result in misdiagnosis.
Otherwise wrongful diagnoses can and do happen sometimes as a result of the constraints of disability benefits, health care and education today. While I am not advocating that a clinician do so, there have certainly been instances in which a wrongful diagnosis has been made for the purpose of gaining disability income for a client, be it workers' compensation, private disability, or government disability. Insofar as health coverage is concerned, a client might have a subclinical condition that the therapist believes would benefit from treatment. But health care coverage might not be provided for anything sub-clinical, so the clinician "bumps up" the diagnosis for the purpose of getting coverage for the client. In education, there are various entitlements and programs that are available only with certain kinds of diagnoses, so a clinician might also provide a diagnosis to a client who could benefit from such entitlements and programs. The constraints of today do seem to promote this sort of misdiagnosis.
At best, diagnosis is far more of an art than it is a science, relying on "evidence" that is quite slippery and hard to assess. Given the difficulties inherent in this process, it is remarkable that we ever get things right.