Optimal Allocation Between a Risky Portfolio and a
Risk-Free Asset
Let an investor allocate a proportion y of total wealth to a risky portfolio P and the
remaining proportion (1−y) to a risk-free asset. Denote the expected return and standard
deviation of the risky portfolio by E(rP ) and σP , respectively, and let the risk-free return
be rf .
1 The Capital Allocation Line
The return on the complete portfolio is
rC = yrP + (1 − y)rf .
Its expected return is
E(rC ) = yE(rP ) + (1 − y)rf = rf + y [E(rP ) − rf ] .
The variance of the complete portfolio is
Var(rC ) = y 2 Var(rP ) + (1 − y)2 Var(rf )
+ 2y(1 − y) Cov(rP , rf ).
Since the risk-free return is constant,
Var(rf ) = 0 and Cov(rP , rf ) = 0.
Thus,
Var(rC ) = y 2 σP2 ,
and, for y ≥ 0,
σC = yσP .
Therefore,
σC
y= .
σP
Substituting this expression into the expected-return equation gives
E(rP ) − rf
E(rC ) = rf + σC .
σP
This is the equation of the capital allocation line. It is a straight line in mean–
standard-deviation space, with
intercept = rf ,
1
and
E(rP ) − rf
slope = .
σP
The slope is the Sharpe ratio of the risky portfolio. It measures the expected excess
return earned per unit of risk.
2 Optimal Portfolio Allocation
Assume that the investor has mean–variance utility:
1
U = E(rC ) − AσC2 ,
2
where A > 0 is the investor’s coefficient of risk aversion, and C denotes the complete
portfolio.
The expected return of the complete portfolio is
E(rC ) = yE(rP ) + (1 − y)rf .
Because the risk-free asset has zero variance and zero covariance with the risky port-
folio,
σC2 = y 2 σP2 .
Therefore, utility can be written as
1
U (y) = yE(rP ) + (1 − y)rf − Ay 2 σP2 .
2
Differentiating with respect to y gives
dU
= E(rP ) − rf − AyσP2 .
dy
The first-order condition for an optimum is
E(rP ) − rf − AyσP2 = 0.
Hence, the optimal proportion invested in the risky portfolio is
E(rP ) − rf
y∗ = .
AσP2
The second-order condition is
d2 U
= −AσP2 < 0,
dy 2
so y ∗ maximizes the investor’s utility.