Consider $P = \binom {-n}k$. This is defined as $$P = {-n\cdot (-n-1)\cdot (-n-2) \cdots (n-(k-2))\cdot(-n-(k-1))\over k\cdot (k-1)\cdots 2\cdot 1}$$ as an analog to the binomial coefficient for positive numbers*. We take $-1$ out of each of the k factors in the numerator to get
$$= (-1)^k\cdot{(n+k-1)(n+k-2)\cdots(n+k-1-(k-2))(n+k-1-(k-1))\over k!}$$
$$P= (-1)^k \binom{n+k-1}k$$
*Note that this is merely an extension of the binomial coefficient/combination to negative numbers, since factorial does not make too much sense for negative numbers. We expand the binomial coefficient to achieve this:
$$n(n-1)(n-2)\cdots(n-(k-1))(n-k)(n-(k+1))(n-(k+2))\cdots 2\cdot 1\over (n-k)(n-k-1)(n-k-2)\cdots 2\cdot 1 \cdot k!$$
$$={n(n-1)(n-2)\cdots (n-k+1)\over k!}$$
This way the processes are equivalent. Indeed, the Maclaurin series of $(1+x)^{-n}, n\in \mathbf N$ would be represented as a polynomial with such coefficients, just like $(1+x)^n$, and in fact these binomial coefficients can be generalised for $n\in\mathbf Q$.