The problem
Let $U_j$ for $0\leq j\leq n$ denote the standard coordinate charts of the complex manifold $\mathbb{P}^n$. Fix $d\geq 1$ and assume we are given holomorphic functions $f_j:U_j\to \mathbb{C}$ for each $j$ with the additionally property that $$f_i=g_{ij}f_j \enspace \text{ on } \enspace U_i\cap U_j $$ For all $0\leq i,j\leq n$, where $g_{ij}=\left(\frac{z_j}{z_i} \right)^d$. Show that there exists a homogeneous polynomial $F$ of degree $d$ such that $$F/z_j^d=f_j \enspace \text{ for all }j. $$
Discussion
The question is one of many problems I am working through to prepare for a graduate level complex analysis exam (no knowledge of vector bundles or the more advanced machinery of manifolds is assumed). We are given a hint to proceed as follows:
Let $G(z_0,...,z_n)$ be a (nonzero) homogeneous polynomial of degree $d$, and set $g_j:=G/z_j^d$ on $U_j$ (for all $j$). Define a map $$h:\mathbb{P}^n\to \mathbb{C}, \enspace h:=\frac{f_j}{g_j} \text{ on } U_j$$ What can be said about $h$?
Let $\pi:\mathbb{C}^{n+1}\setminus \{0\}\to \mathbb{P}^n$ be the natural quotient map. What can be said about $G\cdot (h\circ \pi)$ on $\mathbb{C}^{n+1}\setminus \{0\}$? Why does it extend holomorphically to all of $\mathbb{C}^{n+1}$?
Attempt: Following step $1$, it is easy to argue $h$ is well defined on all of $\mathbb{P}^n$ (if we view it as a map into $(\mathbb{C}\cup \infty)$ )- and it is a meromorphic function on $\mathbb{P}^n$. Let us assume that $h$ could be extended to a holomorphic function on all of $\mathbb{P}^n$ $(*)$.
Then one can show using the maximum modulus principle (and the compactness +connectedness of $\mathbb{P}^n$) that $h$ must be constant. It follows that $F:=G\cdot (h\circ \pi)$ is holomorphic on $\mathbb{C}^{n+1}\setminus \{0\}$, and by Hartog's Theorem can be extended to a holomorphic map on all of $\mathbb{C}^{n+1}$. Since $h$ is constant $F$ is a homogeneous polynomial of degree $d$ (since $G$ is). Moreover, we have (on $U_j$):
$$F([z_0,...,z_n])=G([z_0,...,z_n])\cdot \left(\frac{f_j([z_0,..,z_n])}{G([z_0,...,z_n])} z_j^d\right)=f_j([z_0,...,z_n])z_j^d$$
as desired.
However, this argument relies heavily on the statement $(*)$. Either $h$ can be extended (why would this be true?) or one could choose $G$ more carefully so that $h$ is already holomorphic. Any ideas would be greatly appreciated : )