Discrete-response state space models with conditional heteroscedasticity: An application to forecasting the federal funds rate target

- We propose a discrete-response state space model with conditional heteroscedasticity.
- The proposed model is estimated using MCMC methods.
- The proposed model has better forecast performance than benchmarks that have only constant coefficients or constant variance.

We propose a state space mixed model with stochastic volatility for ordinal-response time series data. For parameter estimation, we design an efficient Markov chain Monte Carlo algorithm. We illustrate our method with an empirical study on the federal funds rate target. The proposed model provides better forecasts than alternative specifications.