The matrix
$$\left( \begin{array}{} -6 & 3 \\ 4 & 5 \end{array} \right)$$
has eigenvalues of -7 and 6. If you turn this matrix upside down,
$$\left( \begin{array}{} 5 & 4 \\ 3 & -6 \end{array} \right)$$
It still has eigenvalues of -7 and 6. Does this happen for any matrix?
Note: I am not asking about a rotation matrix. I'm literally rotating the matrix itself.
Here's an alternate way of stating the problem I came up with while trying to solve this.
You have an arbitrary $n$ by $n$ matrix A. You also have the exchange matrix $J_n$.
$$J_n=\left( \begin{array}{} 0 & 0 & \ldots & 0 & 0 & 1 \\ 0 & 0 & \ldots & 0 & 1 & 0 \\ 0 & 0 & \ldots & 1 & 0 & 0 \\ \vdots & \vdots & & \vdots & \vdots &\vdots \\ 0 & 1 & \ldots & 0 & 0 & 0 \\ 1 & 0 & \ldots & 0 & 0 & 0 \\ \end{array} \right)$$
Then, multiply A by $J_n$ on both sides to get $B$.
$$B = J_n A J_n$$
Do $A$ and $B$ always have the same eigenvalues? If so, why?
Note: I do not care about whether the matrices have the same eigenvectors.
See also: