This paper proposes a probit regression with autocorrelated errors (PAR) to estimate the reaction function of monetary policy in Taiwan using newly constructed binary monetary indicators. We develop a practical sampling scheme via the Gibbs sampling algorithm with data augmentation to make...