The threshold time-varying parameter (TTVP) model

We provide a flexible means of estimating time-varying parameter models in a Bayesian framework. By specifying the state innovations to be characterized by a threshold process that is driven by the absolute size of parameter changes, our model reliably detects whether a given regression coefficient is constant or time-varying at each point in time. Moreover, our modeling framework accounts for model uncertainty in a data-based fashion through Bayesian shrinkage priors on the initial values of the states. In a simulation, we show that our model reliably detects regime shifts in cases where the data generating process displays high, moderate, and low numbers of movements in the regression parameters. Finally, we apply the model to investigate the macroeconomic effects of a US monetary policy shock. We find evidence of pronounced effects prior to the 1970s and less so thereafter. With the outbreak of the global financial crisis, monetary policy effectiveness increases again and even considerably so.