The "pain of a monad" is pretty much the same any many other topics in programming that are tricky for beginners: pointers, recursive functions, linked lists, etc. In fact, monads aren't even a language feature, they are a data structure just like a linked list, defined just like any other data structure. There have even been several articles describing how to implement a monad in other popular languages(monads in ruby, monads in C++, monads in lisp).

As for your second question, it's funny that you ask it after mentioning haskell. The solution of course is that a collection of haskell features offer tools that are "better than Lisp macros" and support side effects. Lazy evaluation and monads cover most cases where you would use a macro in lisp. There are also more complex things like comonads and arrows. These solve the problems that lisp macros solve in a different way. And if you really need the ultimate power that lisp macros provide, then the haskell equivalent is template haskell.