Optimal solutions for optimization in practice 69
In order to determine values for risk aversion, we consider an investor who
holds a portfolio of cash and FTA only. On the basis of the constant abso-
lute risk aversion MV framework, his or her expected utility is expressed by
EU w[()] wμ ( w r)f w
λ
1 22 σ
2
, where w , μ , λ , σ 2 , r f are the optimal pro-
portion in equity, the expected rate of return, the coefficient of absolute risk
aversion, the variance of the FTA, and the riskless rate return, respectively.
Maximizing this utility yields w
rf
μ
λσ^2
or λ
μ
σ
r
w
f
2.^ Considering
current values ( μ 9%, σ 12%, r f 4.6%) reduces this to the trade-off
λ
3
w
. The proportion of equity investment for institutional investors is
indicated by peer-group benchmarks as WM to be 70%. This corresponds to
4 λ 5 as an adequate range for the average institutional investor but prob-
ably inappropriate for high net investors. On the basis of these calculations
within our mean – variance analysis, we obtain λ 2 and λ 10 to represent
high and low risk aversion. These values would fit portfolios of cash and FTA
only with equity positions of 150% and 30%, respectively.
3.5.4 Analysis and comparison
We divide our analysis into four parts, each representing a change in con-
straints, parameters, or the specification of the utility function; the main char-
acteristics are summarized.
Table 3.1 Summary of optimization results: maximum holding 100%
Mean% StdDev% Skewness Kurtosis Gain% Loss% Omega Probability
(a) G/L l 1 12.63 5.73 0.79 3.62 12.68 4.85 2.614 0.72
(b) M-V λ 2 24.40 20.65 0.92 4.41 38.65 19.05 2.029 0.66
(c) M-V λ 10 15.14 8.55 0.90 3.72 18.08 7.74 2.336 0.70
Table 3.2 Summary of optimization results: maximum holding 25%
Mean% StdDev% Skewness Kurtosis Gain% Loss% Omega Probability
(a) G/L l 1 13.66 7.73 0.94 3.86 15.67 6.82 2.298 0.70
(b) M/V λ 2 17.51 11.85 0.96 3.71 23.99 11.27 2.129 0.70
(c) M/V λ 10 15.37 9.37 0.89 3.57 19.38 8.81 2.200 0.67