Investment Management Problem Set 4
Teaching Plan:
1. Discuss briefly questions 1-4, maybe in 5 minutes.
2. Do 7,8,9 .
3. Then do question 10.
4. If time allows do more questions.
1. If markets are efficient, the correlation coefficient between stock returns for two non-
overlapping time periods should be zero. True or false?
True. If not, one could use returns from one period to predict returns in later periods and
make abnormal profits.
2. If markets are semi-strong form efficient, the following are viable strategies to earn
abnormally high trading profits. True or false?
a. Buy shares in companies with low P/E ratios
False.
b. Buy shares in companies with recent above-average price changes
False.
c. Buy shares in companies for which you have private knowledge of an improvement
in management team
True.
In a semi-strong-form efficient market, it is not possible to earn abnormally high profits
by trading on publicly available information. Information about P/E ratios and recent
price changes is publicly known. On the other hand, an investor who has advance
knowledge of management improvements could earn abnormally high trading profits
(unless the market is also strong-form efficient).
3. If the business cycle is predictable, and a stock has a positive beta, the stock’s (abnormal)
returns also must be predictable. True or False?
False. While positive beta stocks respond well to favorable new information about the
economy’s progress through the business cycle, they should not show abnormal returns
around already anticipated events. If a recovery, for example, is already anticipated, the
actual recovery is not news. The stock price should already reflect the coming recovery.
4. Imagine that the fictional business journalist recently argued that “well-managed firms aren't
necessarily more profitable than those with average management." He examined the rates of
return earned by stockholders in a group of 50 firms cited in 2018 as “especially well-
managed", and compared them with another group of stocks that whose management was
average. He found that there was no difference in the stock returns of these two groups of
stocks. He concluded that “If well-managed firms were really more profitable, stockholders'
returns would have been superior".
False. This conclusion is flawed. The main point is that one must draw a distinction between
profitability of a corporation (measured in dollars/pounds—an accounting number), and the stock
return earned by those who hold the shares. If investors know which firms are well-managed,
stockholders will bid up the prices of these shares until the expected return on these securities is
the same as that on the shares of average management firms. This assumes that the quality of
management is not a risk factor, so the required return is the same by investors for both these
companies. But this assumption is not strong, since quality of management is a firm-specific
attribute, and so can be diversified away.
5. Read the paper Edmans, Alex. "Does the stock market fully value intangibles? Employee
satisfaction and equity prices." Journal of Financial economics 101.3 (2011): 621-640.
Summarize its findings, and discuss what they mean for market efficiency.
The paper finds that the 100 companies that are rated as the “best companies to work for”
by the magazine Fortune earn positive abnormal long run returns (alpha=3.5% per year, for
6-7 years after portfolio formation). This means that the information of job satisfaction of
employees does positively affect the future cash flows earned by these firms, but that this
information is not captured by current prices. In other words, high employee satisfaction
firm are undervalued.
The paper makes another important point, that employee satisfaction is intangible. It’s not
something you can observe from the companies balance sheets, like sales growth. The
market does not seem to factoring in the price the fact that more happy employees are more
productive and therefore the companies’ performance is better.
6. The return of Ford is provided by this model rF =0.10% + 1.1rM . If the return of the market is
8%, and the return of Ford is 7%, what is the abnormal return earned by Ford?
The return on the market is 8%. Therefore, the forecast monthly return for Ford is:
0.10% + (1.1 × 8%) = 8.9%
Ford’s actual return was 7%, so the abnormal return was –1.9%.
7. In excel spreadsheet “Lecture 4:Portfolio formed on BE-ME” we calculated the returns to
value strategy, that sells low B/M stocks and buys high B/M stocks. Are the returns we have
documented actually attainable to investors? Discuss.
Those returns assume that the market is perfect, i.e., there are not transaction costs, no short sale
costs, and no taxes. Obviously, these assumptions are not met in the real world. If we were to
actually conduct the strategy, we would have to rebalance the portfolio every year. When doing
so we have to pay transaction costs (via bid-ask spreads) which will erode out returns. More
importantly, in our calculations we short sell the low B/M portfolio, which will entail additional
costs. Finally, on any return we earn we have to pay taxes. So, the answer is that those returns
are not available to investors. The actual return in gross terms would be lower due to transaction
costs, and the take-home return would be even lower due to taxes.
8. The monthly rate of returns on risk-free bonds is 1%. The market went up this month by
1.5%. In addition, AmbChaser Inc., which has a beta of 2, surprisingly just won a lawsuit that
awards it $1M immediately.
a. If the original value of AmbChaser Inc. were $100M, what would you guess was
the rate of return on its stock for that month?
b. What is your answer to (a) above, if the market had expected AmbChaser Inc. to
win $2M in the lawsuit?
Based on broad market trends, the CAPM indicates that AmbChaser stock should have
increased by: 1.0% + 2.0 × (1.5% – 1.0%) = 2.0%
Its firm-specific (nonsystematic) return due to the lawsuit is $1 million per $100 million
initial equity, or 1%. Therefore, the total return should be 3%. (It is assumed here that the
outcome of the lawsuit had a zero expected value.)
If the settlement was expected to be $2 million, then the actual settlement was a “$1 million
disappointment,” and so the firm-specific return would be –1%, for a total return of 2% – 1%
= 1%.
The point here is that we need to differentiate the broad market return given by the CAPM,
form the idiosyncratic returns, which the lawsuit influences.
9. In a recently closely contested lawsuit, Apex sued Bpex for patent infringement. The jury
came back today with its decision. The rate of return on Apex on that day was 3.1%, and of
Bpex was 2.5%. The return on the market that day was 3%. The historical relationship
between the returns of thee stocks and the market is given below:
Apex: 0.2% + 1.4 rM
Bpex: -0.1% + 0.6rM
Based on this information, who do you think won the lawsuit?
Given the model for returns we have, predicted returns on the two stocks would be:
Apex: 0.2% + (1.4 × 3%) = 4.4%
Bpex: –0.1% + (0.6 × 3%) = 1.7%
Apex underperformed this prediction; Bpex outperformed the prediction. We conclude that
Bpex won the lawsuit.
Note that the prediction from a model like this, gives a forecast of returns. The difference
between the actual return and the forecast is the unexpected component, which in this case
is the outcome of the lawsuit.
10. According to the Consumption based CAPM, the following relationship holds:
log Et ( 1+ r M ) −log ( 1+r f )=γcorr (x , r m )σ x σ rm
rm: The return on the market
x: consumption growth
rf: The risk-free rate
γ: The coefficient of risk aversion. Higher γ means higher risk aversion.
[log is the natural logarithm]
a. Discuss the intuition of this equation.
b. Assume that σ rm=16.7 % σ x =0.036 , corr ( x , r m ) =0.4 . Historically, rm is around
7% per year and rf is around 1% per year. What value of γ is required to explain
the observed equity risk premium, according to this model?
c. An individual's wealth will equal either 50,000 or 100,000 each with probability
½ so that expected wealth is E[W] = 75,000. This person, has CRRA utility as
1−γ
W
. Calculate the certainty equivalent that is implied by the coefficient of risk
1−γ
aversion that you have calculated in (b). Is this CE reasonable?
In this model households make investment/consumption/savings decision to maximize their
lifetime utility. In this multi-period version of the CAPM, households want to invest in stocks if
stocks provide a hedge against negative shocks to their consumption, i.e., if stocks have a low
correlation with consumption growth, in those periods that our consumption drops sometimes
stocks will cushion the fall. So stocks provide a hedge.
So, the idea is, the lower the correlation between x and r, the more people like stocks.
The extent to which people like stocks also depends on preference to risk. When people are more
risk averse (higher γ), then the less they will like investing in riskier stocks.
Also note that prices and demand are inversely related. When investors want to invest in stocks
prices go up and expected returns go down. But investors are willing to accept these lower
expected returns, due to the hedge of consumption risk that is achieved by investing in equities.
Acknowledging that corr ( x , r m ) σ x σ rm=Cov ( x , r m ), the model thus make two key predictions:
the equity risk premium increases in both the amount of risk in stocks (covariance of stock
returns with consumption growth) and the risk aversion.
There is also a trade-off between these quantities: if Cov (x , r m ) increases (indicating more risk),
the equity risk premium would remain the same only if γ decreases by some specific amount
(and vice-versa).
To solve the problem in (b) we just need to plug the information in the formula and solve for γ.
ln ( 1.07 )−ln ( 1.01 )=γ∗0.4∗0.167∗0.036
We obtain γ ≈ 24.
Let’s do the calculations to determine the CE for part (c).
Expected value of wealth is 75,000.
−23 −23
50,000 100,000
Expected utility is 0.5 × +0.5 × =−¿1.824E-110 (let’s call this −z)
−23 −23
The certainty equivalent is the amount of money x which if had for sure makes us indifferent
between this amount and the gamble. The more risk averse a person is, the further the CE is
pushed below the expected value of the gamble.
√
−23
CE 23 1
=−¿z. Re-arranging: ≈ £ 51,529.
−23 z × 23
So, a person with γ=24 is indifferent between having £51,529 for sure and the gamble with
EV=75,000.
Note that this CE is barely better than the worst-case outcome of the gamble which is 50,000!
This is very extreme. Even the chance of having wealth that is lower by 1,529 is big enough to
deter this person from taking this gamble, which has a 50% to give you much more (100K). A
person with such a risk aversion will not even leave the house!
The coefficient of risk aversion that is required to explain the equity risk premium is so extreme
and unreasonable, which makes the equity premium puzzle one of the most studied phenomena
in the field of finance.
11. In our pre-recorded material in when we discussed the arbitrage strategy in Excel I asked you
to repeat the exercise by (i) increasing the variance of the macro factor R4 and doing the
calculation again (without idiosyncratic risk). Then try increasing the variance of the
idiosyncratic elements. What did you notice in terms of the performance of the arbitrage
strategy?
Because our position towards the macro factor is perfectly hedged, increasing the variance of the
macro-factor does not affect anything in the strategy. However, when we increase the variance of
the idiosyncratic components, the average cash flow provided by the strategy is similar, but the
variance goes way up. This is because we are not hedged towards these idiosyncratic risks, and
so the variance of the cash flow of the arbitrage strategy, which is proportional to the variance of
these idiosyncratic risks, goes up.
