That may be true in the lower level courses, but around 400 level, the work switches over entirely from calculations to proofs. Proofs are as widely varied and difficult to copy as essays, and are much harder to edit without screwing up unless you really understand them.
As I've said, many students self-enforce (somewhat - it's a continuum), because a Math degree that you didn't earn is barely worth the paper it's printed on.
With most students self-enforcing, the exams can still be rigorous. The 400-level courses can still be rigorous. And nobody will complain that the exams and 400 levels are too hard, because they still learnt enough to deal with them.
It's not a result of Math being "harder to copy". Math is easier to copy than an essay, because it's just not as unique. The difference is cultural. Most Math students don't want to cheat their way through, and so the professors are under no pressure to be soft.
Generally the homework is just meant to give you a good bit of practice, there is no way to pass the class without passing the exams.
Having experienced upper, lower, and graduate level courses in math and engineering (as well as taught at the lower level) I can honestly say that the amount of people that directly copy another student's assignments is much higher than most people think. This type of cheating tapers off as you enter the higher levels. On the other hand, I find that the "helping another student out" or "group study" type of cheating probably occurs with a 75%+ frequency among all levels of academics.
The first type (plagiarism) is definitely cheating while the second type (group think) is most likely classified as cheating but never acted against. The fact that you didn't see any cheating in your Number Theory course could simply be due to the fact that the second type of cheating is much more difficult to spot unless you are intimately familiar with the subject material and solution methods.
A student gets 80% on a 25% weighted "token" homework component, 60% on a 25% "lab" component (due to copying large portions of someone else's report), and 30% on the 50% exam (there's always a few low apples). They pass with 50%.
The lecturer sees that the exam is too hard, and makes it easier the next year.
if your homework is 1/4 of your grade in the class, that class is a "pass me" class. in some math classes the prof refuses to acknowledge that homework actually has a measurable effect on your grade, but he will use its presence to round up instead of down (i.e. 89.5 becomes a 90 instead of an 89)
Cheating doesn't happen when there's virtually no incentive. I stand by my point that you can cheat in Math homework more easily than equivalent non-Math homework, if you want to.
If 400-level Math coures are really impossible to cheat on, but there are 400-level non-Math courses that are possible to cheat on, I'd just argue that this is because they aren't really equivalent.
Still, the exam will be a problem. But if lots of students cheat, the lecturers will start making exams easier, because so many students struggle. And in Australia, there's a big belief amongst the education policy mandarins that exams are useless because they don't teach anything.
Still, most Math majors know their degree is useless if they don't learn anything, and they chose a brutally hard course with no set career path just to learn, so why bother cheating? But Engineering majors is another story.