To me it appears that lhs is some fancy way to count all functions from $[n]$ to $[n]$. I tried several approaches, including interpreting $\binom{n-1}{k-1}$ as number of solutions for $n=x_1+\dots +x_k$ where $\forall_i x_i \ge 1$. But it didn't lend me to anything significant.
I'd appreciate some hints or solutions to this.
That is a consequence of the Abel-Hurwitz formula.
The LHS can be seen as $(n-1)!$ times the coefficient of $x^n$ in the product between $$ \sum_{k\geq 0} k x^k=\frac{x}{(1-x)^2},\qquad \sum_{k\geq 0}\frac{n^k x^k}{k!}=e^{nx}. $$