A New Model of Trend Inflation
This code estimates a new model of trend inflation with 3 features: (1) the trend inflation is constrained to lie in an interval; (2) it allows for a time-varying degree of persistence in the transitory component of inflation; (3) it allows for stochastic volatility.
This nonlinear state space model is estimated using an MCMC algorithm based on fast band matrix routines instead of the Kalman filter.
The methodology is illustrated using an empirical exercise with US CPI inflation. We find the model yields more sensible measures of trend inflation than popular alternatives such as the unobserved components stochastic volatility model. The trend inflation estimates can be downloaded here.