A time-varying parameter structural model of the UK economy

Abstract

We estimate a time-varying parameter structural macroeconomic model of the UK economy, using a Bayesian local likelihood methodology. This enables us to estimate a large open-economy DSGE model over a sample that comprises several di¤erent monetary policy regimes and an incomplete set of data. Our estimation identi…es a gradual shift to a monetary policy regime characterised by an increased responsiveness of policy towards in ‡ation alongside a decrease in the in ‡ation trend down to the two percent target level. The time-varying model also performs remarkably well in forecasting and delivers statistically signi…cant accuracy improvements for most variables and horizons for both point and density forecasts compared to the standard …xed-parameter version.

Decisions in Designing an Australian Macroeconomic Model

“…This raises the interesting issue of how we should go about placing more weight on more recent data when preparing forecastsat the estimation stage, the stage of forecasting residuals or both? Doing so at the estimation stage provides a more unified statistical approach, and Kapetanios et al (2019) show that it improves the accuracy of COMPASS forecasts. However, it would still be possible for equations to run off track under the estimation stage approach because it is designed for 'slowly varying parameter processes' rather than abrupt shifts such as those associated with the global financial crisis.…”

“…When this is not the case, it signals that more work is required on an equation, when time permits, to bring the equation back 'on track'. Kapetanios et al (2019) use an alternative approach to placing more weight on the more recent data for forecasting purposes. Using the Bank of England's DSGE model known as COMPASS, they estimate the model allowing for time-varying parameters.…”

Higher-order dynamic effects of uncertainty risk under thick-tailed stochastic volatility

“…Fortunately, Chib and Ramamurthy ( 2010 ) proposed the random block Metropolis–Hastings (RB-MH) algorithm, which can be used to sample the posterior distribution of high-dimensional DSGE models. Using this method can significantly improve the estimation efficiency of high-dimensional DSGE models containing nonnormal and nonlinear latent variables (Kapetanios et al 2019 ). Therefore, we construct a high-dimensional DSGE-SV- t model that includes the Student’s t -distribution and stochastic volatility feature in shocks.…”

Climate Risks and the Realized Volatility Oil and Gas Prices: Results of an Out-of-Sample Forecasting Experiment

“…In other words, one would need to go beyond the existing models depicting the impact of climate risks on the firstmoment of energy-price returns. While doing this, given relatively weak in-sample, but stronger out-of-sample, forecasting results, theoretical energy economists would need to keep in mind the fact that the structural parameters of the models should in fact be evolving over time, and not constant, along the lines of time-varying Dynamic Stochastic General Equilibrium Models (see for example, [53][54][55]).…”