Advances in Risk Management

272 TIME-VARYING RETURN CORRELATIONS AND PORTFOLIOS

The plots indicate how the GARCH model is able to capture the changes
in volatility associated with the 9/11 incident, whereas the rolling method
tends to smooth out the volatility.
With 20 stocks in the portfolio, there are 180 covariance estimates for
each time period. The average of these correlations calculated using the
rolling model and the DCC model are plotted in Figure 14.2. The rolling

01/03/00

0.0300

0.0400

0.0500

0.0600

0.0700

Std. dev.

0.0800

0.0900

0.1000

06/09/0011/14/0004/21/0109/26/0103/03/0208/08/02

Date

01/13/0306/20/0311/25/0305/01/0410/06/04

GARCH Rolling

Figure 14.1Average volatility-average weekly standard deviations from

3 January 2000 to 27 December 2004

01/03/00

0.2

Date

Correlations

0.22

0.24

0.26

0.28

0.3

0.32

0.34

0.36

0.38

0.4

0.42

0.44

06/09/0011/14/0004/21/0109/26/0103/03/0208/08/0201/13/0306/20/0311/25/0305/01/0410/06/04

DCC Rolling

Figure 14.2Average weekly correlations from 3 January 2000 to

27 December 2004