We investigate and assessment of the effectiveness of Thai monetary transmission mechanism from 4 January 2000 to 30 December 2004. A multivariate asymmetric error correction model is applied to capture the interplay of long-term relationship between the changes between three short-term interest rates for the linear model. On the other hand, using a co-integrated vector auto-regressive Markov switching regime change model where some parameters change according to the economic shock for the non-linear model. We present strong evidence that the linear VECM model fails to capture significant non-linearities in the data generating process. Strong evidence has also been found that all three short-term interest rates have an asymmetric distribution in the intercept term.