I'm trying to find an explicit formula of all the eigenvalues for the following $n$ by $n$ stochastic matrix (sum of each row/column is one):
\begin{bmatrix}0&\frac{1}{n-1}&\frac{1}{n-1}&\cdots&\frac{1}{n-1}\\ \frac{1}{n-1}&\frac{n-2}{n-1}&0&\cdots&0 \\ \frac{1}{n-1}&0&\frac{n-2}{n-1}&0&\cdots \\ \cdots&\cdots&\cdots&\cdots&\cdots \\ \frac{1}{n-1}&0&0&\cdots&\frac{n-2}{n-1} \\ \\ \end{bmatrix}
Since this is a stochastic matrix, it is clear that $1$ is an eigenvalue. After doing some numerical experiments, I believe there are only three distict eigenvalues of this special matrix: $1$, $\frac{-1}{n-1}$, and $\frac{n-2}{n-1}$(with multiplicity $n-2$). I want to show this conclusion formally, but I cannot really decompose the matrix into the sum of identity matrix and zero diagonal matrix, since the first element in the matrix is $0$.
Update: I just realize $e_2-e_i$ will always be an eigenvector with eigenvalue $\frac{n-2}{n-2}$. There are $n-2$ pairs of them so we are done.
The characteristic polynomial of this matrix seems rather straightforward to compute. If my calculations are correct, then we have $$p_A(x)=-x\Big(\frac{n-2}{n-1}-x\Big)^{n-1}-\frac{1}{(n-1)^2}\Big(\frac{n-2}{n-1}-x\Big)^{n-2}+\frac{x}{n-1}\Big(\frac{n-2}{n-1}-x\Big)^{n-3}$$ Can you calculate the eigenvalues from this?