Is using the phrase "... is left as an exercise for the reader" considered good style in academic papers/theses?
I believe it's only appropriate in the type of writing where you are in a position to give the reader excercises. That is, in course books or other learning material. There, I think it's fine, assuming you've already given enough information for them to be able to perform such an excercise.
In papers or theses, you're actually trying to convince the reader that you're right - the reader is in the position of power, not you, as you want something from them (to accept your ideas) and not vice versa. Therefore, I'd never write such a phrase in a paper or thesis.
It's of course a viable requirement to leave out some parts of the work, either because of space constraints or because they're trivial, long but straightforward, would derail the course of the text, or something similar. If possible, such parts can be delegated to an appendix, or left out entirely. But I would accompany this with different phrasing, something like this:
The formula (7) can be obtained from (6) using straightforward application of [insert appropriate math branch here]
A detailed proof of statement (4) unfortunately exceeds the scope of this paper.
The derivation of (3) from (1) is too long to present here, but it can be worked out using a symbolic computation system such as Mathematica.
Equation (7) follows from a straightforward application of this-or-that theorem to equation (5); we refer the reader to existing literature on the topic for details of this.
I will have to advise against it, and very strongly so.
I did applied mathematics in the uni, and some of our teachers were very keen on using phrases like that. Well, I can definitely see the point in leaving exercises to the student, however we did have separate exercise sessions where everything was left as an exercise for the reader. So I have always found it odd (at times frustrating) that the course book, or lectures, by which I was supposed to get the tools to be able to solve the exercises, left some of the necessary bits as an exercise.
There are two major problems with the phrasing, in my humble opinion:
It might not be as obvious or straight-forward as you (the author) might think. In some cases the exercise is practically calculations that have already been performed in the previous page or so, in that case it's understandable to omit the same calculations/derivations, but a reference to the last place these calculations, would do just as well:
"... the rest of the derivations are analogous to..."
or
"... as we have seen in Theorem/Lemma X ... "
That particular phrasing has a condescending/patronizing resonance in the eyes of the reader. This has been something practically all of my old classmates commented on. As a matter of fact, we still joke about the phrase "... is left an exercise to the (ambitious) reader". I am not sure what you have to gain from alienating your audience.
Lastly, regarding the use of the phrase in articles, I have to admit I don't do any work in maths anymore, but in the fields I work with (computational biology, biomedicine, bioinformatics, ...) I have never seen something like this, and I am pretty sure it would not fly well.
If the "exercise" in question is important, you include it in the paper. If you lack the space to include it, you put in supplementary and refer to it. If it is not important you do not mention it.
I don't agree with any of these assessments. Often the omitted deductions are rudimentary, but not necessarily trivial. The reason they are omitted is generally due to lack of space, or the fact that the omitted proofs do not add anything to the exposition, not laziness. The final point doesn't even apply.
That said, stating that something "is left to the reader" is a little cheeky. You often find the equivalent, "an astute reader will be able fill in the details", which may come across as a little pretentious.
For a paper, it is okay to use such a statement, assuming that the author has actually done the proofs.
In a thesis, it is expected that the author does the work, not the reader, so I don't think it is acceptable.