Advances in Risk Management

58 SENSITIVITY ANALYSIS OF PORTFOLIO VOLATILITY

1992

 10

 8

 6

 4

 2

0

2

4

6

Portfolio returns

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

Figure 3.1Portfolio returns

Figures 3.1 and 3.2 depict daily returns of the EWMA optimal portfolio
and the time series of the estimated variance respectively.
In order to assess the importance and impact of weights onσp,we
appliedD. As we are to illustrate, this gives one the possibility of calibrating
strategies with respect to a specific stock or a group of assets, in considera-
tion of both the relative size of the imposed relative weight change [equation
(3.5)].
Table 3.2 shows the SAresults for the considered stocks under the hypoth-
esis of uniform and proportional changes and for the optimal strategy. As
expected, the results imply that the impact of weights on the portfolio
volatility varies depending on the TRS (see also Figure 3.3).
This is also evident from Table 3.3 that shows the ranking of assets and
Figure 3.3 that compares the asset importance in the three strategies.
Let us analyse the ranking in somewhat greater detail. Honeywell Inter-
national Inc. is the most influential stock for uniform and proportional
strategies (for example, it would rank first if one usedPDsorE); it is
the fourth most influential for the optimal strategy. Alcoa In. is the most
important asset in the optimal strategy; it ranks 7th and 5th for uniform
and proportional strategies respectively. Hewlett-Packard Co. is the least
influential weight in the case of uniform changes; it ranks however 3rd for
proportional changes and 7th in the optimal strategy. We note that Table 3.2