As the title suggests, it's not a usual practice but should one do it? All I know that one can study for an exam and simply write the answers needed in order to succeed, no one tells us to include the names of authors of the ideas we memorized to write (except rarely).
In most technical university exams, there is no expectation of citations since the concepts are typically "textbook-level" material and thus established as part of the "common knowledge" of a subject. Such information is rarely cited in the scientific literature, as it taken to be part of the shared background knowledge that has been so accepted as to no longer require explicit crediting. For example, Newton is almost never cited when discussing basic physics.
In a literature or similar humanities course, on the other hand, the subject of an exam may explicitly be discussion and analysis of ideas presented by others. In this case, one would be typically be expected to at least informally cite the people whose ideas are being discussed, e.g.:
Rawls' theory of justice, while more sophisticated than the simple utilitarianism of Locke, is considerably more difficult to comprehend"
Full formal citation, however, is typically both impractical and not useful within the scope of an exam.
Where business courses fall on this spectrum will depend on the course: when discussing case studies, I would expect informal citation to be appropriate, whereas an operational research and statistics course would more likely fall into the technical end of the spectrum where there are no citations.
Proper citation? No. I couldn't even give you a citation of one of my own papers from memory.
Should you be able to recall a name of some prominent figure that is relevant to your exam? Probably, but that depends on the type of material your professor expects to learn. I've taken plenty of exams like this.
Building on jakebeal's excellent answer, let's consider that citation has two purposes:
- Let the reader know what ideas and findings are your own original contributions.
- Inform readers where they can find the stuff that doesn't belong to you.
As others have mentioned here, #2 is practically infeasible in a standard exam venue, at least to the standard of most academic writing. Do you have a photographic memory of every article and book you have read, including page numbers and full bibliographic data (I think Smith and Johnson found that x > 3 when p < 5, but was that C Smith and D Johnson, or D Smith and C Johnson? Or maybe it was C John Smith and D Jones? Was this in the May issue or the April issue?)? What remains, then is to inform your readers that you are basing part of your answer on someone else's work, quite possibly in the form of a disclaimer. So, you could say something like:
According to a study in the late 1970's, quantum hypotunneling was found to have a statistically significant benefit in trimming the positive ion matrix when the average manifold area under the curve is finite. As I have shown above (using Jackson's Stepwise Theorem) that this area with respect to the problem at hand is, in fact, finite, we can conclude that quantum hypotunneling is likely an effective solution. McFielding's theory that the positive ion matrix is a purely sociopolitical construct and ineffective for practical engineering work does not apply for the following reasons....
Perhaps we should consider that the exam is just a piece of a larger work - the course itself - which already has a reference section: the syllabus.
If during the exam you feel compelled to use material from outside the assigned reading of the course, then you should cite it in such a way that the marker can easily find it.
Or perhaps you should confine yourself to the assigned reading.