Central bank intervention typically entails the use of multiple and possibly non-linear policies. In this paper we introduce a dynamic Tobit model embedded in a Vector Autoregression in order to identify simultaneous monetary shocks. Our method is easily estimated using only least squares and a maximum likelihood function. Also, impulse-responses are carried out as in the traditional time-series setting and can be applied in a structural framework. In simulation exercises we show that, as policy covariance grows, our method increasingly outperforms a benchmark case of estimating policy functions separately. We find significant differences when estimating our method in emerging market economies.