Tutorial 10 Solutions – Portfolio Management – II

[Readings: Ch18]

Q1 Discuss the key features of the following portfolio performance measures:

• Sharpe Ratio
• Treynor Ratio
• Jenson’s Alpha
• The Information Ratio

Ri − RFR
Sharpe Ratio measure: Si = where i represents total risk.
i

• Sharpe uses standard deviation of returns as the measure of risk - portfolio

performance is measured in relation to the CML

• Treynor measure uses beta (systematic risk) - examines portfolio performance based
on the SML.

• Sharpe therefore evaluates the portfolio manager on the basis of both rate of return
performance and diversification
• Produce relative not absolute rankings of performance
• The methods agree on rankings of completely diversified portfolios

Treynor Ratio recognised two components of risk

– Risk from general market fluctuations
– Risk from unique fluctuations in the securities in the portfolio
His measure of risk-adjusted performance focuses on the portfolio’s nondiversifiable risk,
market or systematic risk

Ri − RFR
T=
i
The numerator is the risk premium (average return on security i over time less the
average risk free rate).
The denominator is a measure of systematic risk
The expression, T, is the risk premium return per unit of risk
Risk averse investors prefer to maximize this value
This assumes a completely diversified portfolio leaving systematic risk as the relevant
risk.

1
Jensen Alpha
• Based on CAPM
• Expected return on any security or portfolio is:
E ( R j ) = RFR +  j [ E ( RM ) − RFR]
Where: E(Rj) = the expected return on security j
RFR = the one-period risk-free interest rate
j= the systematic risk for security or portfolio j
E(Rm) = the expected return on the market portfolio of risky assets

E(Rj) and RFR change over time, hence a time series of rates is employed.The empirical
CAPM includes an error term, and in risk-premium format it is:

E ( R jt ) − RFRt =  j [ E ( Rmt ) − RFRt ] + e jt

Superior portfolio managers earn higher risk premiums than those suggested by the
model. Superior performance may be measured by adding an intercept term (alpha) to
the equation as follows:

E ( R jt ) − RFRt =  j +  j [ E ( Rmt ) − RFRt ] + e jt

j indicates if the manager is superior ( >0) or inferior ( <0).
Some advantages of Jenson’s measure over Treynor & Sharpe:

• easier to interpret ~ the value of alpha indicates how good the manager is

• alpha is empirically estimated using regression, enabling tests of significance to

be applied

• Jenson’s measure is flexible enough to allow for alternative models – single factor
(such as CAPM) & multifactor (such as Fama & French 3-factor model).

The Information Ratio measures average return in excess of benchmark portfolio

divided by the standard deviation of this excess return

R j − Rb ER j
IR j = =
 ER  ER
where ER is the residual (unsystematic) risk, also called the tracking error.

• The Sharpe ratio may be seen as a special case of the IR where the risk free asset is
assumed to be the benchmark portfolio
• If excess portfolio returns are estimated with historical data using the same single-
factor regression equation used to compute Jenson’s alpha, the IR simplifies to:
j
IR j =
 e where e is the standard error of the regression.

2
Q2 Describe two major factors that a portfolio manager should consider before designing
an investment strategy. What types of decisions can a manager make to achieve these
goals?

The two major factors would be: (1) attempt to derive risk-adjusted returns that exceed a
naive buy-and-hold policy and (2) completely diversify - i.e., eliminate all unsystematic risk
from the portfolio. A portfolio manager can do one or both of two things to derive superior
risk-adjusted returns. The first is to have superior timing regarding market cycles and adjust
your portfolio accordingly. Alternatively, one can consistently select undervalued stocks.
As long as you do not make major mistakes with the rest of the portfolio, these actions should
result in superior risk-adjusted returns.

Q3 An analyst wants to evaluate Portfolio X, consisting entirely of U.S. common stocks,

using both the Treynor and Sharpe measures of portfolio performance. The following
table provides the average annual rate of return for portfolio X, the market portfolio (as
measured by the S&P 500), and the U.S. Treasury bills during the past eight years:
Average annual Standard deviation Beta
Rate of return of return
Portfolio X 10% 18% 0.6
S&P 500 12% 13% 1.0
T-bills 6% N/A N/A

(i) Calculate the Treynor and Sharpe measures for both Portfolio X and the S&P
500. Briefly explain whether Portfolio X under-performed, equalled, or out-
performed the S&P 500 on a risk-adjusted basis using both Treynor and
Sharpe measures.

(ii) Based on the performance of Portfolio X relative to the S&P 500 calculated in
part (i), briefly explain the reason for the conflicting results when using the
Treynor measure versus the Sharpe measure.

(i) Treynor measure:

S&P500: (12 – 6)/1 = 6

Portfolio X: (10 – 6)/0.6 = 6.67

(ii) Sharpe measure:

S&P500: (12 – 6)/13 = 0.462

Portfolio X: (10 – 6)/18 = 0.222

Portfolio X outperforms the S&P500 based on the Treynor measure, but

underperforms based on the Sharpe measure. The inconsistency is due to differences
in the unit risk measure. Treynor measure uses beta risk, while Sharpe measure is a
total risk measure. This implies that Portfolio X has some unsystematic risk in its total
risk (i.e. it is not well diversified). If Portfolio X were well diversified, then both
measures will provide similar rankings.

3
Q4 Consider the following table:

Average Standard
Return % p.a. Deviation % p.a. Beta
Market Index 10.7 19.0 1.00
Fund I 19.4 44.2 1.34
Fund II 13.5 21.8 1.12
Fund III 12.9 17.6 1.30
Fund IV 9.3 5.6 0.71

a) If the return on long term Commonwealth bonds is 5.6% per annum,

calculate the following performance measures for each of the funds in the
table and the market index.

(i) Jensen's alpha P = R P − R f + P ( R M − R f )

SP =
(R − Rf )
10.7 − 5.6
=
= 0.268
(ii) Sharpe index
P

P 19.0
R − R f 10.7 − 5.6
(iii) Treynor index TP = P = = 5.10
P 1

Jensen's alpha Sharpe Index Treynor Index

Market Index - 0.268 5.100

Fund I 6.966% 0.312 10.299
Fund II 2.188% 0.362 7.054
Fund III 0.670% 0.415 5.615
Fund IV 0.079% 0.661 5.211

b) Compare and comment on your answers for part a).

From part a), the performance measures provide different rankings. Under
Jensen’s alpha, Fund I is clearly preferred and the ranking is Funds I, II, III and
IV. Under the Sharpe index, all funds perform better than the market benchmark
and the ranking of funds is reversed, ie. Funds IV, III, II and I. Under the Treynor
index, the ranking is yet again reversed with Fund I now the highest ranked, ie.
Funds I, II, III and IV. The results show that inconsistency in rankings is possible
under different performance measures.

c) Which fund has exhibited the best performance?

Given the inconsistency in rankings, it is difficult to claim one fund as superior.
Fund I is ranked best under Jensen’s alpha and the Treynor index while Fund IV
is ranked best under the Sharpe index. Both Jensen’s alpha and the Treynor
index use beta as the risk measure while the Sharpe index uses standard
deviation as the risk measure. Hence, the ranking of funds depends upon the
investor’s assessment as to the appropriate risk measure (which may include
consideration of the level of diversification).

4
 Q5

(i) Explain the circumstances in which the Sharpe and Treynor indices can provide
conflicting fund rankings.

The Sharpe and Treynor measures can provide conflicting rankings. The
inconsistency is due to differences in the unit risk measure. The Sharpe Index uses
standard deviation whereas the Treynor Index uses beta risk. However, if the fund
is well diversified then nonsystematic risk will be largely eliminated and the Sharpe
and Treynor Indices will provide very similar rankings.

(ii) What is the Carhart (1997) model? How does it differ from the Jensen’s alpha
measure?

Carhart’s (1997) empirical model is expressed as:

Rp – Rf = αp + βpRm +spSMB +hpHML +upUMD +εp

Carhart’s model has four return generating factors: 1) Rm is the return on the market
index; 2) SMB is the return on the mimicking size portfolio; 3) HML is the return on the
mimicking book-to-market portfolio; and 4) UMD is the return on the mimicking
momentum portfolio. Carhart’s model is supposedly used to control for market biases.

Now if we assume that Carhart’s model is an accurate one, then the error returns can be
assumed away, allowing the estimation of alpha for any portfolio:

αp = (Rp – Rf) – (βpRm +spSMB +hpHML +upUMD)

Carhart’s alpha, as described above, is a measure of superior performance after controlling for
the forces generated by the market return, size premium, value premium and especially
momentum premium (many funds utilise the momentum strategy in their portfolio
management).. Hence, any fund managers that construct portfolios designed to capture these
premiums will find that their returns are captured within the model, and so they will not
exhibit any alpha performance. Rather, managers that have strategies that do not follow
mainstream premiums are expected to have alpha performance. In this sense, Carhart’s alpha
is regarded in the industry as a more appropriate measure of individual performance. Jensen’s
alpha, on the other hand, assumes that the CAPM is the appropriate benchmark, to the extent
that the CAPM is valid. Hence, Jensen’s alpha relies upon an estimate of beta that may be
problematic. Further, the measure is claimed to only measure depth and not breadth. For
instance, a fund manager may have invested in a large number of stocks on which several
losses have been made, but also picked a stock on which a substantial profit was made. The
resultant value of Jensen’s alpha could be positive, but this is due to one lucky winner and not
consistent performance across the portfolio. In such a case, it could be incorrectly concluded
that the manager had superior skills.

5
(iii) A substantial amount of evidence on the performance of funds has provided
inconsistent results. What are some examples of this inconsistency? What are some of the
explanations for the inconsistency?

Early studies in this area typically found that funds, on average, under-performed the
benchmarks. The conclusion from studies was strong evidence against the ability of fund
managers to out-perform market benchmarks. Subsequent studies, especially in the USA revealed
inconsistent results. Some studies found evidence of positive alphas indicating superior fund
performance. But the results are sensitive to the sample, time-period and the market index.
Specifically, the use a value-weighted market index appears to result in lower values of Jensen’s
alpha than an equal-weighted market index.

Research during the 1980s focussed on more specific performance attributes such as market
timing, consistent with the advancement in the theoretical performance models. This research
has generally shown that fund managers do possess market timing ability but again the results
are mixed. The likely explanation for the conflicting results is in different samples over different
time periods and different performance benchmarks. Data from different funds and over different
time periods indicates sensitivity in the results. Moreover, depending on the selected
performance benchmark, different funds and indeed funds in aggregate perform differently.

Q6 Consider the following information regarding the performance of a money

manager in recent month. The table represents the actual return of each sector of
the manager’s portfolio in column 1, the fraction of the portfolio allocated to each
sector in column 2, the benchmark or neutral sector allocations in column 3, and
the returns of sector indices in column 4.

Actual Actual Benchmark Index

return weight Weight return
Equity 2% 0.7 0.6 2.5% (S&P 500)
Bonds 1% 0.2 0.3 1.2% (Salomon Brothers index)
Cash 0.5% 0.1 0.1 0.5%

(i) What was the manager’s return in the month? What was her over-performance
or under-performance?
(ii) What was the contribution of security selection to relative performance?
(iii) What was the contribution of asset allocation to relative performance? Confirm
that the sum of selection and allocation contribution equals her total “excess”
return relative to the bogey.
(iv) Which investment decisions would you allow this manager to make?

6
 (i) Benchmark: 0.6(2.5%) + 0.3(1.2%) + 0.1(0.5%) = 1.91%
Actual: 0.7(2.0%) + 0.2(1.0%) + 0.1(0.5%) = 1.65%
Underperformance - 0.26%

(ii) Security selection:

(1) (2) (1)x(2) = (3)
Market Excess return Manager’s portfolio Contribution
(Manager – index) weight to performance

Equity -0.5% 0.70 -0.35

Bond -0.2% 0.20 -0.04
Cash 0 0.10 0

Contribution of security selection -0.39%

(iii)
Asset allocation:
(1) (2) (1)x(2) = (3)
Market Excess weight Index return Contribution
(Manager – benchmark) - Index Overall to performance

Equity 0.10 0.59 0.059

Bond -0.10 -0.71 0.071
Cash 0 -1.14 0

Contribution of asset allocation 0.13%

Summary:
Security selection -0.39%
Asset allocation 0.13%
Excess performance -0.26%

(iv) This manager should only be allowed to make equity and bond asset allocation
decisions.

Applied Discussion: Discuss the main points in the following article: [BHB, pp.685-686]

In the news
Fund persistence
Unfortunately, fund manager performance, like bands, can be a bit hit and miss. State
Street Global Advisors (SSgA) has studied the performance of Australian fund managers
and concluded that their persistence, that is, their ability to consistently add value and
achieve above-benchmark investment returns, is rather too Korgiesque for comfort. A
Number One hit today is no guarantee of further Number Ones in future. The analysis,
using wholesale fund performance data sourced from William M Mercer, suggests most
active Australian equity managers have a tough time producing ‘persistently above-
benchmark returns after fees’.

7
 Susan Darroch, head of structured products for SSgA in Australia, says investors
need to be aware of whether their manager can consistently outperform, ‘as most do not
with any measure of persistence’. ‘There’s always a question of whether active managers
can outperform,’ Darroch says. ‘We wanted to get the results that Mercer has and have a
look to see if we could prove or disprove that active managers can outperform. That’s all
we were doing, just seeing whether, on average, they outperformed, and whether they
could continue to outperform.
‘The conclusion is that, as a whole, the average active manager finds it difficult to
continually persist in outperforming.’
…
Since 2001 most managers have performed below the threshold. Sixty managers
achieved index-plus-2.0 per cent over a three-year period, but only 23 of those 60 (38 per
cent) duplicated their success in the subsequent three years’ and that the ‘median
international equity manager showed very slight outperformance from 2000 to 2002.
More recent results have been slightly less awesome: managers successfully achieved
index plus 2.0 per cent 118 times over a three-year period, but only half duplicated their
success in the subsequent three years.
In terms of picking a good funds manager, Peter Gunning, chief investment officer
Asia Pacific for Russell Investment Group, states that ‘investors should expect periods
when any active fund manager underperforms not only a market index or benchmark but
its peers as well. It’s a simple, logical and understandable result of the methods fund
managers use to manage money. For a manager to make money for investors, its approach
has suit what’s going on in the market. From time to time different factors drive the
performance of investment markets, and a manager’s style may not work as well under
some conditions as under others.
…
‘So there is a lot more to it than past performance,’ Gunning says. ‘But it’s a logical
for investors to use past performance to pick managers, because often it’s the only
information they have available. But that does not mean it’s the best information to make
a decision on.’
Source: Simon Hoyle, ‘All you need is a ticket to ride’, Sydney Morning Herald, 22 October
2005.

Commentary
The above article examines the twin issues of performance (addressed in Chapter 19) and the
relevance of past performance or persistence, discussed earlier in this chapter. To summarise,
outperforming the benchmark return, such as the market index return, is difficult and that ‘as a
whole, the average active manager finds it difficult to continually persist in outperforming’. In
picking a good manager for futures years, investors should look at the investment strategy and
resources available to the manager before making a decision. As Peter Gunning attests, ‘From
time to time different factors drive the performance of investment markets, and a manager’s style
may not work as well under some conditions as under others.’ However, for the novice investor,
this information may be difficult to obtain and understand. As discussed in the chapter, a natural
consequence is that investors may demonstrate an overreliance on past information, such as past
performance.