64 SENSITIVITY ANALYSIS OF PORTFOLIO VOLATILITYThe additive property of theDis then useful in calibrating TRS aim-
ing at diversification across sectors. In our simple 30-stock portfolio, let us
consider five types of industries: manufacturing, financial services, telecom-
munication and information and communication technologies, energy,
and others (as a residual category). As Table 3.1 shows, stocks have
been divided in those groups by following a simple criterion over the
corresponding company main area of activity. Of course, the categoriza-
tion may result as loose or not adequate in some cases, but it has the
advantages of considering a small number of industries and it does fit
our scope of showing the empirical application of theDto a portfolio
GARCH(1,1).
Table 3.4 shows theDfor the five sectors. By additivity, they are the
sum of theDof the assets belonging to the group, over different strategies.
Figure 3.4 presents the absolute values ofDin Table 3.4.
It is interesting to note that:
Manufacturing is the most important sector for the uniform and optimal
strategies. However, it is among the least influential (ranking 4th) for a
proportional strategy;
Telecommunication and ICT is the most influential sector if a proportional
strategy is assumed; it ranks second for the uniform strategy and least
for the proportional one;
Financial services assets rank 3rd, 5th and 4th respectively; hence they
tend to have a low impact with respect to the change in portfolio volatility
in the various strategies.
Energy assets rank 5th, 3rd and 2nd in the three strategies: their influence
varies significantly across trading/reallocation strategies, as they are the
least influential for a uniform strategy, while ranking 2nd for the optimal
strategy;
Others assets rank 4th, 2nd and 3rd for the three strategies respectively:
they tend to maintain an intermediate relevance across the strategies.
A final note, when recalling the definition ofDand the interpretation of
volatility, the above results can be readily interpreted in terms of risk. For
example, considering the optimal strategy, most of the diversification would
be due to the Manufacturing and Energy categories, since the corresponding
assets are responsible for 66 percent of the change inσp.
3.6 CONCLUSION
In this study we have illustrated the sensitivity analysis (SA) of portfolio
volatility (σp) estimation models. We have suggested that performing the