Novel approaches to portfolio construction 43
Interestingly, such an exercise not only helps the PM discover portfolios with
better transfer coefficient but also gives him or her specific guidelines, namely
the combination of tracking error and turnover values, about how to modify
his or her strategy to attain them.
Furthermore , once the PM has demarcated the universe of desirable portfo-
lios by overlaying the heatmaps of expected return and transfer coefficients,
he or she can move on to examine heatmaps of other metrics such as implied
beta ( Figure 2.13 ), ratio of specific and tracking error ( Figure 2.14 ), etc., thus
zooming into the universe of portfolios that are of interest to him or her.
Case study
We conclude this section by discussing an illustrative example. Consider a
PM trying to rebalance his or her portfolio using the following strategy.
Maximize expected return
s .t
● Budget constraint
● Tracking error constraint (at most 2%)
● Turnover constraint (at most 10%)
● Asset, industry, sector, style bounds constraint
● Active beta constraint (at most 1%)
● Threshold holding and trading constraint
● Average daily volume holding and trading constraintTable 2.3 gives the key characteristics of the starting portfolio. Given the
souring economic conditions, the PM is inclined toward pursuing investment
options with as little tracking error as possible, even if that entails violating
some of the other constraints by a small amount. Furthermore, portfolios with
higher transfer coefficient are deemed extremely desirable. Since the account
under consideration is taxable, the PM is particularly interested in tax-efficient
investment options although he or she admits that minimizing tax liability is to
be used only as a secondary objective with maximizing expected return being
the primary one.
Table 2.4 gives the key characteristics of the optimal portfolio, referred to
as P0 in the sequel, obtained by rebalancing the starting portfolio using the
strategy mentioned above. Note that the tracking error bound of 2% is not
attained, whereas the entire turnover budget of 10% is utilized. P0 has an
expected return of 0.2794% and a transfer coefficient of 0.6391. Next, we dis-
cuss a series of steps that the PM can take to improve P0. The flowchart shown
in Figure 2.15 gives a brief summary of what follows.