Suppose that $f: \emptyset \to \emptyset$ is the empty function. The only possible inverse function, I believe, is that for which there are no elements in the domain or codomain, and the statement $(f \circ f^{-1})(x) = x$ for all $x$, and vice versa, holds vacuously.
Is this correct?