The problem is that this approach requires patience, persistence and hard work. Do this help students in the long-term? If you learn math without struggling, will you learn to think in the same way? Is this teaching to the test over teaching you how to think?
The struggle mostly comes from situations where a) people just can't follow the explanations at all, since the reasoning is not explained well enough, or there are jumps too large, or b) people who have trouble adapting past strategies and concepts to more novel situations. For example, they might know how to solve some class A of word problems, but once you switch the word problems to solving the equation in the other direction, they just get lost.
I think most of the 'left as an exercise for the reader' stuff appears once you get to a high enough level that you a) assume that the reader is proficient enough and b) that the reader actually cares enough about the subject to be able to do so. For example, my friends taking a bunch of pure math subjects get that crap all the time. "So we've proved X theorem for Y case, Z will be left as an exercise", and they eat it up (partly because it really is an exercise).
Now, I still think a lot of the time, it's inappropriate, and just used cause the writer is lazy, or has used up too much space on diagrams (especially in physics textbooks...).
There are actual exercises, where solving it is just a matter of applying some concept. However, a lot of math books have things where you have beat your head against the wall to figure them out. Although, the internet has made it easier to lookup, nowadays.
School textbooks often try to make things simpler. However, when these texts are written based off of older texts, which skip steps, they may inherit the style of the older works.