FIE400E - Investments

Topic 3 - Capital Allocation

Francisco Santos

Norwegian School of Economics

Outline

Portfolio of one risky asset and a risk-free asset.

Expected value, variance, and covariance of portfolios.
Capital allocation line.
Diversification.
Systematic versus idiosyncratic risk.
Portfolio of two risky assets.
Portfolio of two risky assets and a risk-free asset.

Readings: BKM, chapters 6-7

Francisco Santos (NHH) FIE400E - Investments 1 / 48

Capital Allocation Across Risky and Risk-Free Portfolios

How much of the portfolio to place in risk-free assets versus other

risky asset classes?

Denote
I P as the investor’s portfolio of risky assets .

I F as the risk-free asset.

Assume P comprises two mutual funds, one invested in stocks (S)

and one invested in long-term bonds (B).

Francisco Santos (NHH) FIE400E - Investments 2 / 48

Capital Allocation Across Risky and Risk-Free Portfolios
Example

Asset Investment Weights I Weights II Weights III

Risk-Free – F 90.000 NOK 30% 30%
Risky Portf. – P 210.000 NOK 70%
Stocks – S 113.400 NOK 37,8% 54%
Bonds – B 96.600 NOK 32,2% 46%

Francisco Santos (NHH) FIE400E - Investments 3 / 48

Portfolio of One Risky Asset and a Risk-Free Asset

For now, we will assume that the investor has already decided the
composition of the risky portfolio.

The investor’s problem is to choose how much to invest in the risky

portfolio P and in the risk-free asset F .

Francisco Santos (NHH) FIE400E - Investments 4 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Denote:
rP : rate of return of P

E [rP ]: expected rate of return of P

σP : standard deviation of P

rF : rate of return of F

C : complete portfolio

rC : rate of return of C

σC : standard deviation of C

y : proportion invested in the risky portfolio P

Francisco Santos (NHH) FIE400E - Investments 5 / 48

Statistical Interlude

Consider:
I X , Y and W : random variables

I a, b, and c: coefficients
P
µX = E [X ] = p(s)X (s)
s

σX2 = Var [X ] = s p(s)[X (s) − E (X )]2

P

P
σX ,Y = Cov [X , Y ] = s p(s)[X (s) − E (X )][Y (s) − E (Y )]
Cov [X ,Y ]
ρXY = Corr [X , Y ] = σX σY

Cov [X , X ] = Var [X ]

Francisco Santos (NHH) FIE400E - Investments 6 / 48

Statistical Interlude

Consider:
I Z = aX + bY

I T = aX − bY

I U = aX + bY − cW

E [Z ] = E [aX + bY ] = E [aX ] + E [bY ] = aE [X ] + bE [Y ]

E [T ] = E [aX − bY ] = E [aX ] + E [−bY ] = aE [X ] − bE [Y ]

E [U] = E [aX + bY − cW ] = E [aX ] + E [bY ] + E [−cW ] = aE [X ] + bE [Y ] − cE [W ]

Francisco Santos (NHH) FIE400E - Investments 7 / 48

Statistical Interlude
Consider:
I Z = aX + bY

I T = aX − bY

I U = aX + bY − cW

Var [Z ] = Var [aX + bY ] = Var [aX ] + Var [bY ] + 2Cov (aX , bY )

= a2 σX2 + b 2 σY2 + 2abσX ,Y

Var [T ] = Var [aX − bY ] = Var [aX ] + Var [−bY ] − 2Cov (aX , bY )

= a2 σX2 + b 2 σY2 − 2abσX ,Y

Var [U] = Var [(aX + bY ) − cW ]

= Var [aX + bY ] + Var [−cW ] − 2Cov (aX + bY , cW )
= a2 σX2 + b 2 σY2 + 2abσX ,Y + c 2 σW
2
− 2acσX ,W − 2bcσY ,W

Francisco Santos (NHH) FIE400E - Investments 8 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
We can determine rC :

rC = yrP + (1 − y )rF

Taking expectations:

E [rC ] = E [yrP + (1 − y )rF ]

= yE [rP ] + (1 − y )rF
= rF + y [E (rP ) − rF ]

Numerical example:

E [rP ] = 15% σP = 22% rF = 7%

E [rC ] = 7 + y (15 − 7)
Francisco Santos (NHH) FIE400E - Investments 9 / 48
Portfolio of One Risky Asset and a Risk-Free Asset

E [rC ] = rF + y [E (rP ) − rF ]

Interpretation:
The base return of any portfolio is the risk-free rate.

In addition, the portfolio is expected to provide a risk premium.

The risk premium depends:

I on the risk premium of the risky portfolio: E (rP ) − rF ;

I on the investor’s position on the risky asset: y .

Francisco Santos (NHH) FIE400E - Investments 10 / 48

Portfolio of One Risky Asset and a Risk-Free Asset

What about σC ?

Var [rC ] = Var [yrP + (1 − y )rF ]

= Var [yrP ] + Var [(1 − y )rF ] + 2Cov [yrP , (1 − y )rF ]
= y 2 Var [rP ]
p q
σC = Var [rC ] = y 2 Var [rP ] = y σP

The standard deviation of the complete portfolio is proportional to

the standard deviation of the risky portfolio.

Francisco Santos (NHH) FIE400E - Investments 11 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
We can combine E [rC ] and σC :

E [rC ] = rF + y [E (rP ) − rF ]
σC = y σP
Solving σC = y σP in terms of y and then plugging into E [rC ], we get:

[E (rP ) − rF ]
E [rC ] = rF + σC
σP

In our example:

8
E [rC ] = 7 + σC
22

Francisco Santos (NHH) FIE400E - Investments 12 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Capital Allocation Line – CAL
The expected return of the complete portfolio as a function of its
standard deviation is a straight line with:
I Intercept = rF
[E (rP )−rF ]
I Slope = σP = Sharpe ratio

In our example:

Francisco Santos (NHH) FIE400E - Investments 13 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Capital Allocation Line – CAL

It depicts all the risk-return combinations available to investors.

The slope of the CAL is the Sharpe ratio.

Every portfolio in the CAL achieves the same Sharpe ratio.

Expected return and standard deviation as function of y (fraction

invested in the risky portfolio):
y E [rC ] σC
0 E [rC ] = 7% σC = 0%
0,5 E [rC ] = 7% + 0, 5 × 8% = 11% σC = 0, 5 × 22% = 11%
1 E [rC ] = 7% + 1 × 8% = 15% σC = 1 × 22% = 22%
1,5 E [rC ] = 7% + 1, 5 × 8% = 19% σC = 1, 5 × 22% = 33%

Francisco Santos (NHH) FIE400E - Investments 14 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Capital Allocation Line – CAL

How can you achieve y = 1, 5?

Leveraging!

value invested in the risky portfolio

y=
total value of portfolio
Suppose you initial investment budget is 300.000 NOK.

You borrow 150.000 NOK.

You invest everything in the risky portfolio.

300.000 + 150.000
y= = 1, 5
300.000

Francisco Santos (NHH) FIE400E - Investments 15 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Capital Allocation Line – CAL
To be more realistic:
I “normal” investors cannot borrow at the risk-free rate.
I Hence, the borrowing cost will exceed the risk free-rate.
I Suppose the borrowing cost is 9% (rFBorrow = 9%).
I This has an implication for the CAL: if y > 1 the Sharpe ratio is a little
bit lower due
( to higher cost of borrowing.
15−7 8
I Sharpe = 15−9 22 = 22 , if y ≤ 1
6
22 = 22 , if y > 1

Francisco Santos (NHH) FIE400E - Investments 16 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Risk-Aversion

Which portfolio in the CAL to choose?

We need to address risk-aversion!

We will assume an utility function:

U = E [r ] − 0, 5Aσ 2

We know the expected return and standard deviation of a portfolio

consisting of the risk-free asset and a risky portfolio:

E [rC ] = rF + y [E (rP ) − rF ]
σC = y σP

Investors attempt to maximize their utility by choosing y .

Francisco Santos (NHH) FIE400E - Investments 17 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Solution to the Problem

Investors’ problem:

max U = E [rC ] − 0, 5AσC2 = rF + y [E (rP ) − rF ] − 0, 5Ay 2 σP2

y

First order condition (F.O.C.):

E (rP ) − rF
E (rP ) − rF − Ay ∗ σP2 = 0 ⇔ y ∗ =
AσP2
maximise utility
45
First highest sharp ratio then max utility

Francisco Santos (NHH) FIE400E - Investments 18 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
U= 2 E(r) - 0,5*A*o^2
Solution to the Problem y* = (2 E(r) - rf)) / (Ao^2)

U = E(r)(1+0,5o^2)^A
y* = you need to derive
=> E(r) - 0.5*A*o^2 ->
U = E(r) - 2o^2
A = 4 --> 0,5*4 = 2
max. utility
Using our numerical example and an investor’s risk aversion of 4:

15% − 7%
y∗ = = 41%
4 × 0, 222
This yields:
I E [rC ] = 7% + 0, 41(15% − 7%) = 10, 28%

I σC = 0, 41 × 22% = 9, 02%

Francisco Santos (NHH) FIE400E - Investments 19 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Indifference Curve Analysis

A graphical way of presenting this decision problem is to use

indifference curve analysis.

Consider an investor with a risk-aversion A = 4.

Consider that this investor currently holds all her wealth in a risk free
asset yielding rF = 5%.

This is a risk-free asset, so the utility score U = 5%.

We have to find the expected return that the investor will demand for
holding some risk.

Francisco Santos (NHH) FIE400E - Investments 20 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Indifference Curve Analysis

Imagine that an investor moves to a portfolio with σC = 1%.

For the investor to be indifferent between this portfolio C and the

risk-free asset, C should yield the same utility U = 5%.

We need to find E [rC ]:

U = E [r ] − 0, 5Aσ 2

0, 05 = E [rC ] − 0, 5 × 4 × 0, 012 ⇔ E [rC ] = 5, 02%

This investor is indifferent between holding the risk-free asset or a

risky portfolio C with σC = 1% and E [rC ] = 5, 02%.

Francisco Santos (NHH) FIE400E - Investments 21 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Indifference Curve Analysis

Steeper curves mean that investors require a greater increase in

expected return to compensate for an increase in risk.

Francisco Santos (NHH) FIE400E - Investments 22 / 48

Portfolio of One Risky Asset and a Risk-Free Asset
Indifference Curve Analysis

Francisco Santos (NHH) FIE400E - Investments 23 / 48

Portfolio of One Risky Asset and a Risk-Free Asset

Special case of the CAL: the Capital Market Line (CML)

The CML is the CAL provided by 1-month T-bills and a broad index
of common stocks.

Francisco Santos (NHH) FIE400E - Investments 24 / 48

Diversification and Portfolio Risk

Suppose your portfolio is composed of only one stock.

What are the sources of risk to this portfolio?

I Risk that comes from conditions in the general economy.

I Risk that comes from firm-specific conditions.

What happens if you add another stock to your portfolio?

I The portfolio is still subject to risk coming from conditions in the
economy.

I But, to the extent that firm-specific influences on the two stock differ,
diversification should reduce portfolio risk.

Francisco Santos (NHH) FIE400E - Investments 25 / 48

Diversification and Portfolio Risk

When common sources of risk affect all firms, diversification can

reduce the portfolio’s risk, but not fully eliminate it.

Market risk: risk arising from market wide risk sources.

I Cannot be diversified away.

I Also called systematic risk, or nondiversifiable risk.

Firm-specific risk: risk arising from firm-specific conditions.

I Can be eliminated by diversification.

I Also called unique risk, nonsystematic risk, diversifiable risk, or

idiosyncratic risk.

Francisco Santos (NHH) FIE400E - Investments 26 / 48

Diversification and Portfolio Risk

Francisco Santos (NHH) FIE400E - Investments 27 / 48

Portfolio of Two Risky Assets

Consider two mutual funds:

I One specializing in bonds – B;

I One specializing in stocks – S.

Consider the following information about the mutual funds:

B S
E [r ] 8% 13%
σ 12% 20%
Cov (rB , rS ) 0,0072
Corr (rB , rS ) 0,3

Francisco Santos (NHH) FIE400E - Investments 28 / 48

Portfolio of Two Risky Assets

Suppose you invest the following proportions in the funds:

I wB in the bond fund;

I wS in the stock fund.

Thus, you have a portfolio P = wB B + wS S.

The rate of return of this portfolio is: rP = wB rB + wS rS .

The expected return of this portfolio is: E [rP ] = wB E [rB ] + wS E [rS ].

The variance of this portfolio is: σP2 = wB2 σB2 + wS2 σS2 + 2wB wS σB,S .

Francisco Santos (NHH) FIE400E - Investments 29 / 48

Portfolio of Two Risky Assets

The variance of this portfolio is: σP2 = wB2 σB2 + wS2 σS2 + 2wB wS σB,S .

It is clear that the variance of the portfolio is reduced if the

covariance term is negative.

But, it is even better than that!

Unless B and S are perfectly positively correlated (ρB,S = 1), the

portfolio standard deviation is less than the weighted average of the
individual security standard deviations1 .

This is important, because the expected return of the portfolio is the

weighted average of the individual expected returns.

Hence, we gain in terms of Sharpe ratio.

1
Proof in next slide.
Francisco Santos (NHH) FIE400E - Investments 30 / 48
Portfolio of Two Risky Assets
Proof:
We start with σP2 = wB2 σB2 + wS2 σS2 + 2wB wS σB,S .

We know that: σB,S = ρBS σB σS

Substituting:

σP2 = wB2 σB2 + wS2 σS2 + 2wB wS ρBS σB σS

Considering the extreme case of ρB,S = 1. In that case:

σP2 = wB2 σB2 + wS2 σS2 + 2wB wS σB σS

= (wB σB + wS σS )2

Thus:

σP ≤ wB σB + wS σS with equality if ρBS = 1

Francisco Santos (NHH) FIE400E - Investments 31 / 48

Portfolio of Two Risky Assets

This has very important implications:

Portfolios of less than perfectly positively correlated assets always
offer better risk-return opportunities than the individual component
securities on their own.

The lower the correlation between the assets, the greater the gain in
efficiency.

Francisco Santos (NHH) FIE400E - Investments 32 / 48

Portfolio of Two Risky Assets
Expected Return
We can compute the expected return for different compositions of our
portfolio P.

This is not a function of ρS,B .

Francisco Santos (NHH) FIE400E - Investments 33 / 48

Portfolio of Two Risky Assets
Standard Deviation
We can compute the standard deviation return for different
compositions of our portfolio P.

This is function of ρS,B .

Francisco Santos (NHH) FIE400E - Investments 34 / 48

Portfolio of Two Risky Assets
Investable Combinations
Joining the two previous figures:

Francisco Santos (NHH) FIE400E - Investments 35 / 48

Portfolio of Two Risky Assets
Minimum-Variance Portfolio – MVP
There is a portfolio that minimizes risk – the Minimum-Variance
Portfolio (MVP).
We can determine the MVP by solving:
min σP2 = wB2 σB2 + wS2 σS2 + 2wB wS σB,S
wB ,wS

Given that wB + wS = 1 we can instead solve:

min σP2 = wB2 σB2 + (1 − wB )2 σS2 + 2wB (1 − wB )σB,S
wB

F.O.C.
2wB∗ σB2 − 2(1 − wB∗ )σS2 + 2(1 − 2wB∗ )σB,S = 0
This yields:
σS2 − σB,S
wB∗ =
σS2 + σB2 − 2σB,S
wS∗ = 1 − wB∗
Francisco Santos (NHH) FIE400E - Investments 36 / 48
Portfolio of Two Risky Assets
Solution to the Problem – Optimal Portfolio

Note that the MVP minimizes risk, but that does not imply that is
the best portfolio to invest.
For that, we need again to address risk aversion.
We need to maximize investor’s utility:

max U = E [r ] − 0, 5Aσ 2
wB ,wS

= wB E [rB ] + wS E [rS ] − 0, 5A[wB2 σB2 + wS2 σS2 + 2wB wS σB,S ]

Solving it, yields the weights for the optimal portfolio of two risky
assets:
E [rB ] − E [rS ] + A(σS2 − σB,S )
wB∗ =
A(σB2 + σS2 − 2σB,S )
wS∗ = 1 − wB∗

Francisco Santos (NHH) FIE400E - Investments 37 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Consider now that the investor can choose between a risk-free asset
and the two mutual funds.

Suppose rF = 5% and ρB,S = 0, 3.

For each feasible portfolio, we can draw a CAL.

Consider first the CAL that goes through the minimum-variance

portfolio.

Applying the formulas to our previous example of bonds and stocks:

0,22 −0,0072
I wB∗ = 0,122 +0,22 −2×0,0072 = 0, 82

I wS∗ = 1 − 0, 82 = 0, 18

I E [rMVP ] = 0, 82 × 8% + 0, 18 × 13% = 8, 9%
p
I σMVP = 0, 822 × 0, 122 + 0, 182 × 0, 22 + 2 × 0, 82 × 0, 18 × 0, 0072 =
11, 45%
Francisco Santos (NHH) FIE400E - Investments 38 / 48
Portfolio of Two Risky Assets and a Risk Free Asset
The Sharpe ratio of the CAL (slope) that passes through the
minimum-variance portfolio is:
E [rMVP ]−rF 8,9−5
I SMVP = σMVP = 11,45 = 0, 34

Francisco Santos (NHH) FIE400E - Investments 39 / 48

Portfolio of Two Risky Assets and a Risk Free Asset

Consider now a CAL that passes trough a portfolio that invests 70%
in bonds and 30% in stocks – portfolio Z.

E [rZ ] = 0, 7 × 8% + 0, 3 × 13% = 9, 5%
p
σZ = 0, 72 × 0, 122 + 0, 32 × 0, 22 + 2 × 0, 7 × 0, 3 × 0, 0072 = 11, 70%

E [rZ ]−rF 9,5−5

SZ = σZ = 11,70 = 0, 38

Francisco Santos (NHH) FIE400E - Investments 40 / 48

Portfolio of Two Risky Assets and a Risk Free Asset

CAL(Z) dominates CAL(MVP), but why stop at Z?

Francisco Santos (NHH) FIE400E - Investments 41 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Optimal Risky Portfolio

We should maximize the Sharpe ratio:

E [rP ] − rF
max SP =
wi σP
X
subject to wi = 1
i

With two risky assets, the weights of the optimal risky portfolio are:

E [RB ]σS2 − E [RS ]σB,S

wB∗ =
E [RS ]σB2 + E [RB ]σS2 − (E [RS ] + E [RB ])σB,S
wS∗ = 1 − wB∗

where RB = rB − rF and RS = rS − rF

Francisco Santos (NHH) FIE400E - Investments 42 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Optimal Risky Portfolio

In our example:
(0,08−0,05)0,22 −(0,13−0,05)0,0072
wB∗ = (0,08−0,05)0,22 +(0,13−0,05)0,122 −(0,08−0,05+0,13−0,05)0,0072
= 40%

wS∗ = 1 − 0, 4 = 60%

E [r ] = 0, 4 × 8% + 0, 6 × 13% = 11%
p
σ= 0, 42 × 0, 122 + 0, 62 × 0, 22 + 2 × 0, 4 × 0, 6 × 0, 0072 = 14, 20%

11−5
S= 14,20 = 0, 42

Francisco Santos (NHH) FIE400E - Investments 43 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Optimal Risky Portfolio

The portfolios in CAL(TP) provide the highest Sharpe Ratio.

To choose a specific one, we need to address risk aversion.
Francisco Santos (NHH) FIE400E - Investments 44 / 48
Portfolio of Two Risky Assets and a Risk Free Asset
Solution to the Problem

Consider an investor with a risk aversion coefficient A = 4.

We can use the results from slide 18 , where P is now replaced by TP:

E (rTP ) − rF
y∗ = 2
AσTP

11% − 5%
y∗ = = 74, 39%
4 × 0, 1422

Francisco Santos (NHH) FIE400E - Investments 45 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Investor’s Optimal Portfolio

This investor will invest:

25.61% in the risk-free asset;
74,39% in the tangent portfolio;

Tangent portfolio consists of 40% in bonds and 60% stocks.

Overall weight of the bond fund is wB = 40% × 74, 39% = 29, 76%.
Overall weight of the stock fund is wS = 60% × 74, 39% = 44, 63%.

E [r ] = 0, 2561 × 5% + 0, 2976 × 8% + 0, 4463 × 13% = 9, 46%.

p
σ = 0, 29762 × 0, 122 + 0, 44632 × 0, 22 + 2 × 0, 2976 × 0, 4463 × 0, 0072 =
10, 56%.
9,46−5
S= 10,56
= 0, 42

Francisco Santos (NHH) FIE400E - Investments 46 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Investor’s Optimal Portfolio

Francisco Santos (NHH) FIE400E - Investments 47 / 48

Portfolio of Two Risky Assets and a Risk Free Asset
Investor’s Optimal Portfolio

Francisco Santos (NHH) FIE400E - Investments 48 / 48