Presentation is loading. Please wait.

Presentation is loading. Please wait.

Simulation techniques Martin Ellison University of Warwick and CEPR Bank of England, December 2005.

Similar presentations


Presentation on theme: "Simulation techniques Martin Ellison University of Warwick and CEPR Bank of England, December 2005."— Presentation transcript:

1 Simulation techniques Martin Ellison University of Warwick and CEPR Bank of England, December 2005

2 Baseline DSGE model Recursive structure makes model easy to simulate

3 Numerical simulations Stylised facts Impulse response functions Forecast error variance decomposition

4 Stylised facts Variances Covariances/correlations Autocovariances/autocorrelations Cross-correlations at leads and lags

5 Recursive simulation 1. Start from steady-state value w 0 = 0 2. Draw shocks {v t } from normal distribution 3. Simulate {w t } from {v t } recursively using

6 Recursive simulation 4. Calculate {y t } from {w t } using 5. Calculate desired stylised facts, ignoring first few observations

7 Variances Standard deviation Interest rate0.46 Output gap1.39 Inflation0.46

8 Correlations Interest rate Output gap Inflation Interest rate1 Output gap11 Inflation11

9 Autocorrelations t,t-1t,t-2t,t-3t,t-4 Interest rate0.500.250.120.06 Output gap0.500.250.120.06 Inflation0.500.250.120.06

10 Cross-correlations Correlation with output gap at time t t-2t-1tt+1t+2 Output gap0.250.501 0.25 Inflation0.250.501 0.25 Interest rate-0.25-0.50-0.50-0.25

11 Impulse response functions What is effect of 1 standard deviation shock in any element of v t on variables w t and y t ? 1. Start from steady-state value w 0 = 0 2. Define shock of interest

12 Impulse response functions 3. Simulate {w t } from {v t } recursively using 4. Calculate impulse response {y t } from {w t } using

13 Response to v t shock

14 Forecast error variance decomposition (FEVD) Imagine you make a forecast for the output gap for next h periods Because of shocks, you will make forecast errors What proportion of errors are due to each shock at different horizons? FEVD is a simple transform of impulse response functions

15 FEVD calculation Define impulse response function of output gap to each shocks v 1 and v 2 response to v 1 response to v 2 response at horizons 1 to 8

16 Shock Impulse response at horizon 1 Contribution to variance at horizon 1 FEVD at horizon h = 1 At horizon h = 1, two sources of forecast errors

17 FEVD at horizon h = 1 Contribution of v 1

18 Shock Impulse response at horizon 2 Contribution to variance at horizon 2 FEVD at horizon h = 2 At horizon h = 2, four sources of forecast errors

19 FEVD at horizon h = 2 Contribution of v 1

20 FEVD at horizon h Contribution of v 1 At horizon h, 2h sources of forecast errors

21 FEVD for output gap

22 FEVD for inflation

23 FEVD for interest rates

24 Next steps Models with multiple shocks Taylor rules Optimal Taylor rules


Download ppt "Simulation techniques Martin Ellison University of Warwick and CEPR Bank of England, December 2005."

Similar presentations


Ads by Google