A lot of questions about this have unsatisfying answers that either argues how unsafe RSA is when $p=q$ or points out that $\phi(n) \neq (p-1)(q-1)$ for $p=q$.
However, I'd like to know why the RSA fails, i.e why the Decipher $R$ is not equal to the Message $M$ when $p=q$, regardless of safety concerns and assuming the public and private keys are just blindly computed in modulo $(p-1)(q-1)$.