2004 CFA L 3 – S tudy T ips Schweser CFA Level 3 Study Tips 2004 Test Format As you no doubt are aware, your Level 3 examination will be 50 percent essay and 50 percent multiple choice (item set), with the essay portion in the morning and the multiple choice in the afternoon. Since AIMR® doesn’t release old multiple choice questions, it’s very difficult to predict what you’ll see in the afternoon session. However, we have provided many old AIMR® essay questions, as well as new ones we’ve written, to help you prepare. You’ll find these questions in our study notes. Your 2004 Level 3 Study Guide states that 10 percent of the exam will be Ethics and Professional Standards, 0 to 10 percent will be quantitative analysis, 30 to 40 percent asset valuation (Equity, Debt, and Derivatives), and 40 to 60 percent Portfolio Management. Remember that AIMR reserves the right to test material in either essay or multiple choice format. For example, the new material on behavioral investing is quite suitable as a portion of an essay or multiple choice question. Even Ethics, which is an odds-on favorite for item set format, could show up as part of an essay. Content There is a lot of new material this year, much of it in derivatives. If I were making my CFA®-season study plan (I’ve included a sample 18-week plan in book one of our study notes), I would definitely plan extra time at the end to go back and re-study that material. Also, the behavioral/psychological investing material isn’t particularly difficult, but I would read it thoroughly and be familiar with it. Study Session 18 (global investing) is new this year, but it doesn’t really add much to the curriculum. Much of it was already covered in material that was removed from other study sessions for 2004. Strategy DO NOT take Level 3 lightly! I strongly recommend that you make a study plan and follow it throughout the season. And don’t deviate from your plan assuming you can “make it up” in the coming weeks. I realize this is very tempting, but I speak from experience in saying it’s very difficult to make up for lost time! 1
36
Embed
Study Session 2 – Quantitative Analysis · Web viewThey could have a financial advisor saying something like, “the stock isn’t expected to move much, so let’s put together
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
2004 CFA L 3 – S tudy T ips
Schweser CFA Level 3 Study Tips 2004
Test Format As you no doubt are aware, your Level 3 examination will be 50 percent essay and 50 percent multiple choice
(item set), with the essay portion in the morning and the multiple choice in the afternoon. Since AIMR®
doesn’t release old multiple choice questions, it’s very difficult to predict what you’ll see in the afternoon
session. However, we have provided many old AIMR® essay questions, as well as new ones we’ve written,
to help you prepare. You’ll find these questions in our study notes.
Your 2004 Level 3 Study Guide states that 10 percent of the exam will be Ethics and Professional Standards,
0 to 10 percent will be quantitative analysis, 30 to 40 percent asset valuation (Equity, Debt, and Derivatives),
and 40 to 60 percent Portfolio Management. Remember that AIMR reserves the right to test material in either
essay or multiple choice format. For example, the new material on behavioral investing is quite suitable as a
portion of an essay or multiple choice question. Even Ethics, which is an odds-on favorite for item set format,
could show up as part of an essay.
ContentThere is a lot of new material this year, much of it in derivatives. If I were making my CFA®-season study
plan (I’ve included a sample 18-week plan in book one of our study notes), I would definitely plan extra time
at the end to go back and re-study that material. Also, the behavioral/psychological investing material isn’t
particularly difficult, but I would read it thoroughly and be familiar with it. Study Session 18 (global investing) is
new this year, but it doesn’t really add much to the curriculum. Much of it was already covered in material that
was removed from other study sessions for 2004.
StrategyDO NOT take Level 3 lightly! I strongly recommend that you make a study plan and follow it throughout the
season. And don’t deviate from your plan assuming you can “make it up” in the coming weeks. I realize this is
very tempting, but I speak from experience in saying it’s very difficult to make up for lost time!
If at all possible, maintain a steady exercise program. Try to eat right and get plenty of sleep, especially in the
6 – 8 weeks before the exam. Keeping yourself in good shape will help keep your mind and body sharp under
the stress of the exam.
Quantitative Methods: Study Session 3 Quantitative methods is always a sure thing for the exam. Look for a regression output that you will have to
utilize/analyze. You will probably be asked to forecast the value of a dependent variable using the
coefficients, as well as evaluate the significance of the individual coefficients and the overall model. I strongly
1
2004 CFA L 3 – S tudy T ips
feel that you will be asked to “fill in the blanks” in an ANOVA output. Here is a sample regression output
including the ANOVA output with the relationships of the various cells, so you can get a start on feeling
comfortable with the process:
(The model is trying to estimate a home building company’s annual sales as a function of GDP and changing
interest rates.)
df = degrees of freedom; SS = sum of squares; MSS = mean sum of squares; n = total observations;
So, the manager started with T-bills valued at $120,000,000 and ended up with a total value of
110,750(ST), the value of the stocks in the index.
He thus earns the return on the stocks in the index.
Portfolio Management – Individual Investors: Study Session 9Portfolio Management, individual and institutional, is without question the most important material for the
Level 3 exam. Portfolio management will be 50 to 60 percent of the exam and will dominate the morning
(essay) portion of the exam. This week I will focus on individual portfolio management.
I strongly recommend that you read and reread our study notes and try as many old exam questions as
possible. Learning the patterns in answering AIMR® essay questions is paramount. By the way, as you do
the old exam questions, be aware that the guideline answers are not what AIMR expects for the exam.
Typically, you can use about half of what AIMR includes in their guideline answers and get full credit. As you
read their answer, give thought to how much you could leave out.
Here are a few suggestions for answering essay questions:
Get to the point. Be as brief as possible while fully answering the question.
Don’t feel you have to say a lot to say enough
NEVER add anything that does not relate directly to the question asked.
19
2004 CFA L 3 – S tudy T ips
Don’t give the grader excuses to deduct points by making unnecessary comments or
assumptions.
Even if AIMR gives you a full page for a one-line answer, just write the one line!
Determine the time horizon(s) first.
The various horizons will be clear from the case.
Working years (pre-retirement).
Intermediate goals, such as college for kids.
Retirement.
Post-retirement.
Given the time horizons and their required cash flows, determine the required return.
Consistency, consistency, consistency.
Be sure your final recommendation agrees with your goals and constraints.
Even if you recommend the wrong portfolio (by choosing one from among four or five), if
you have been consistent throughout the answer, you will get some points.
If you are not consistent and recommend the wrong portfolio, you will miss all the points.
If the client’s willingness to accept risk (as stated in the case) does not agree with his or her
ability to accept risk (as determined by wealth, income, goals, and time horizon), go with
the client’s desires unless doing so will place the portfolio at undue risk.
If you recommend less risk, be sure to make a note in your answer to explain why
you are not taking as much risk as the client wants.
It is almost always safe to recommend risk/return education.
You can usually recommend legal counsel, too, especially if there are any unusual
circumstances or complications. Examples include special trusts set up at before
or after death, large charitable contributions, etc.
Make logical assumptions.
Inflation of 3% is reasonable.
Protection of principal is reasonable (i.e., don’t meet cash flows out of principal,
unless specifically stated in the case). For example, the client may want to make
an immediate donation to charity. The donation must be deducted from the
portfolio before determining time horizons and associated returns/cash flows
Behavioral Investing
Look for “The Psychology of Investing” (beginning on page 63, Book 4) to show up in cases for individual
investors. (I wouldn’t be surprised to see item set questions, also!) It would be very simple for AIMR to have
the client make certain statements that demonstrate overconfidence, snake bite, fear of regret, house money,
etc. If you detect any psychological traits that place constraints on the portfolio that are not in the client’s best
interest, be ready to point them out and explain why the client needs education on the topic.
20
2004 CFA L 3 – S tudy T ips
Portfolio Management – Institutional Investors (Study Session 10)This material, which focuses primarily on employer-sponsored pension plans and insurance companies, has
been in the curriculum for a while and is usually tested in essay format. Having said that, however, be aware
that it would be very easy for AIMR® to test pensions, both defined benefit and defined contribution, in the
item set format. If pension plans show up on the exam, I would expect calculations and explanations (essay
format) to be more relevant for defined benefit pension funds. Defined contribution plans would more
appropriately be tested as item set, since the major institutional problems associated with them typically
center on the firm’s fiduciary duty to the employees (beneficiaries).
Look over thoroughly and be able to answer questions related to the summary table on page 122 of Book 4.
Don’t just memorize the table, however. Remember, Level 3 is all about being able to look at the material
from one perspective and answer from another. For example, for a defined benefit pension plan, the table
states the amount of risk tolerance, “depends upon surplus, age of work force, time horizon, and company
balance sheet.”
Be able to explain each component of the statement. For example, “The risk tolerance of the defined benefit
pension plan depends upon the average age of the work force, because average age determines the time
horizon and the associated pension liability. High average age implies a short time horizon and high pension
liability. A short time horizon usually means low risk tolerance, especially when accompanied by a large
liability. ”
Surplus: The larger the surplus the more risk the plan can take. However, a safety net (minimum) return
should be determined. If the return on the fund’s assets hits the safety net return, sometimes referred to as a
trigger return, the fund manager should immunize. (Remember contingent immunization? It is discussed on
page 76 of Book 2.)
Age of work force: Generally, the younger the work force, the longer until they retire, and an older average
age work force indicates a shorter horizon. How do surplus and work force age relate? The younger the work
force, the longer the planning horizon, and the smaller the necessary surplus (i.e., the more time to make up
for short term underperformance).
Time horizon: Time horizon is an extremely important variable to pinpoint. Without first determining the time
horizon, meaningful required and safety net returns are impossible to determine with any validity.
Company balance sheet: By now you’ve noticed the circularity of the characteristics already discussed. Any
pension fund planning means next to nothing if the company is in financial difficulty. Of course, given a fairly
sound firm, the stronger the balance sheet (the greater the equity of the firm) the more risk that can be taken,
21
2004 CFA L 3 – S tudy T ips
because the firm is better able to make special contributions to the fund. If the firm is unable to do that, the
pension fund manager must be very careful not to accept undue risk in the plan’s assets.
Comparing Institutional and Individual Investors
Other than the obvious differences associated with size and the firm’s fiduciary duty to its employees, legal
and time horizon issues are important differences for the institutional investor versus the individual investor.
For example, the time horizon for the institutional investor is usually infinite with continuing liabilities to current
retirees while the individual’s time horizon is finite, because even if the individual intends to leave money to
charity or relatives, the investor has a finite life.
Strategies for Asset Allocation (Study Session 11)The summary points on page 165 are very important for the exam. Again, don’t just memorize the points;
know what they mean. For example, take the statement:
“A constant mix strategy will outperform buy and hold and CPPI strategies in a flat but oscillating market.”
The buy and hold strategy does nothing in a flat, oscillating market, so the value of the portfolio oscillates
with the market, but neither transactions costs nor profits or losses associated with trading are incurred.
CPPI assumes readjustment of the portfolio according to a formula. As the portfolio value falls (oscillates
downward) and approaches the floor value, the portfolio manager sells equities. As such the portfolio
manager sells low. Then, as the market oscillates back up, the portfolio manager will buy equities at
increased prices (buys high).
In a constant mix strategy, the manager will continually buy low and sell high. As the portfolio increases in
value, due to increases in the value of equities held, the manager will be forced to sell equities to maintain the
constant percentage mix of debt and equity. Then, as the market falls and the percentage of equities in the
portfolio falls with it, the manager must buy equities to return to the set percentage.
Bottom line, when the market oscillates but is generally flat, the buy and hold strategy will experience no
transactions costs, profits, or losses associated with trading. In the CPPI, the manager buys high and sells
low, continually generating losses. In the constant mix, the manager buys low and sells high, continually
generating profits. Hence the constant mix strategy outperforms both the buy and hold and CPPI strategies in
a flat, oscillating market.
Portfolio Management: The Tracking Error Question (Study Session 14)From Study Session 14:
Tracking error = standard deviation of monthly excess returns.
22
2004 CFA L 3 – S tudy T ips
Excess return = alpha = the difference between the portfolio return and the index (benchmark) return.
Earlier in the curriculum:
Tracking error = the difference between the portfolio return and the index return.
Tracking error risk = the standard deviation of the tracking error over time.
The confusion:
Tracking error = the difference between the portfolio and index returns (i.e., a single measurement at one
point in time).
Tracking error = the standard deviation of the difference between the portfolio and index returns over time
(i.e., the standard deviation of many measurements).
Alpha = excess return = the difference between the portfolio and index returns, (i.e., a single measurement at
one point in time). This means alpha is the same as one of the tracking error definitions!
For the exam:
If there is a silver lining to this dark cloud, it is that AIMR® must realize the confusion as well. On the exam
they will have to be very explicit in what they ask. For example:
Bounty Funds has experienced a tracking error of 28% over the past twelve months, and the
average excess return (alpha) during the period was 6%. Calculate Bounty’s information ratio.
The information ratio is defined as:
IR = (average excess return, or alpha) / (standard deviation of excess returns) = 0.06/0.28 = 0.214
In this example, we know they must be using tracking error defined as the standard deviation of
excess returns over the period, because it was presented as a measure over several periods.
Using the data provided in Figure 1, calculate Bounty Fund’s average tracking error over the last
twelve months.
Figure 1: Benchmark and Portfolio Returns
MonthReturn
Tracking ErrorPortfolio Benchmark
1 0.010 0.009 +0.001
2 0.008 0.002 +0.006
3 0.014 0.011 +0.003
23
2004 CFA L 3 – S tudy T ips
4 0.006 0.010 –0.004
5 –0.001 0.006 –0.007
6 0.005 0.007 –0.002
7 0.002 –0.009 +0.011
8 0.011 0.010 +0.001
9 0.007 0.007 +0.000
10 0.010 0.006 +0.004
11 0.001 –0.001 +0.002
12 0.002 0.008 –0.006
Average = +0.00075
By the way the data was presented and the question was asked, we know they calculate tracking error as the
simple difference between portfolio and benchmark returns and they want the average of that difference for
the past twelve months.
Using the data in Figure 1 (and assuming a standard deviation of 0.009), calculate the information
ratio for Bounty Funds.
IR = (average excess return, or alpha) / (standard deviation of excess returns) =0.00075/0.009 =
0.0833
The question was phrased such that you know exactly what you need to do (i.e., what form of
tracking error was to be used). In this question tracking error was presented as the standard
deviation of excess returns.
Risk Considerations: VAR (Study Session 15)The two methods for calculating value at risk (VAR) are analytical and historical:
Analytical VAR
Using the analytical method to express VAR in percentage terms:
24
2004 CFA L 3 – S tudy T ips
Note: On page 18 of Book 5, currency VAR is presented as:
By multiplying by the value of the portfolio, this equation turns the percentage VAR into dollars (currency).
IMPORTANT NOTE: Both forms of the equation are found in the CFA curriculum this year. The first equation,
however, subtracts the absolute (positive) value of z times the portfolio standard deviation, and the second
adds the actual (negative) value of z times the portfolio standard deviation. Both expressions, therefore,
subtract z(s) from the expected return.
Analytical VAR is based upon past variance and covariance measures (i.e., the analyst must assume past
variances and covariances are representative of the future). Also, since you are only concerned with
downside risk, VAR is a one-tail measure. As such, we do not necessarily work with the z-values we used to
create confidence intervals (CI).
In the back of Book 5 are two tables: the cumulative z-table and the alternative z-table. The alternative z-
table, which could also be used to generate VAR, is what you used to generate confidence intervals. The
cumulative z-table, however, is actually better for measuring VAR, because it measures one tail rather than
two. All you have to do is find the level of significance you desire in the body of the table.
Aside: The graphic at the top of the table might confuse you, because it seems to measure the upper tail.
Since the distribution is symmetrical, however, the z-values actually measure either tail.
Figure 1 has a few VAR significance levels and their corresponding z-values from the cumulative z-table. Be
sure you are confident in finding these values:
25
2004 CFA L 3 – S tudy T ips
Once you have located the desired significance value (actually as close to it as possible), you add the
corresponding values for z along the left margin and across the top. Looking for 0.9000, for example, you will
find a value of 0.8997. We add the corresponding values for z and get 1.2 (row heading) + 0.08 (column
heading) = 1.28. This means we will use z = 1.28 to define the 90 percent VAR.
Try it with a 95 percent VAR: 95 percent implies 5 percent in the tail or a value in the body of the table of
0.9500. Searching through the table, 0.9505 is about as close as you can get, so z = 1.6 + 0.05 = 1.65.
Example:
Let’s assume we want a 90 percent annual VAR for a portfolio with an expected return of 15 percent and a
standard deviation of 9 percent. Since we want 10 percent of the possible returns to fall in the lower tail, the
z-value is 1.28, and:
If we now assume the portfolio is worth $1,000,000,
There are two interpretations of these answers:
There is 95% probability the annual return will be greater than 0.0348 or $34,800.
There is 5% probability of an annual return less than 0.0348 or $34,800.
Note: Figure 2 shows z-values using the alternative z-table, just in case that is your preference. You will
notice that, in each case, the values in the table are those in Figure 1 less 0.5000.
26
2004 CFA L 3 – S tudy T ips
Historical VAR
Historical VAR would also be easy to ask on the exam. Historical VAR is based upon actual past returns, not
their distribution; hence, it is not dependent upon past measures of variance and covariance.
Example:
Estimate the 90 percent 1-week VAR for Bounty Funds. Figure 3 presents the past 20 weekly returns for
Bounty Funds, in no particular order:
Answer:
First, we rank the returns in descending order from the highest 1-week return: 0.0032, 0.0028, 0.0021,
Ninety percent VAR implies 90 percent or 0.90(20) = 18 of the returns are above the 90 percent VAR and
0.10(20) = 2 returns are below. From the ranking, we see that the two lowest returns are –0.21 percent and –
0.27 percent.
We can say two things:
There is 90% probability that the lowest 1-week return will be –0.21%.
There is 10% probability that the 1-week return will be lower than –0.21%.
Note: VAR must always have a time dimension (e.g., in our first example the VAR was stated for a 1-year
period, and in the second, VAR was stated for a 1-week period). VAR must also be stated with a set
percentage. We used 90 percent in both the examples. Although you might see any value used on the exam,
I would expect that you will have to calculate a 95 percent VAR. From Figure 1, you can see that z = 1.65 is
associated with a 95 percent VAR.
27
2004 CFA L 3 – S tudy T ips
Performance Evaluation, Attribution, and GIPS® (Study Session 17)This material is part of the 40–60 percent allocation to portfolio management. You will see a performance
presentation that you will have to analyze—either a prepared presentation or statements related to its
construction.
GIPS
Every year AIMR® has used at least one question related to performance presentation. Be sure to focus on
old exam questions as an indicator of what you will see in this area on the exam. Know the requirements—
especially the disclosures. In recent years, you needed to identify errors in presentation along with omissions.
The omissions almost always relate to missing required disclosures. See page 197, Book 5, for a complete
list and discussion of required disclosures.
Start with our coverage of GIPS in Study Session 17 (Book 5) and the example on page 219. Next, go to the
problem set associated with our Study Session 17 review, which starts on page 225. Once you feel fairly
comfortable with the material, work through the old exam questions, starting on page 236 with 2002 Question
4.
Performance Evaluation
The Sharpe and Treynor measures and Jensen’s Alpha, starting on page 101 (in Study Session 16,
Evaluation of Portfolio Performance) in Book 5, are always important topics. The Sharpe measure, for
example, has been on almost every Level 3 exam in recent history, and candidates have been asked to
compare evaluation results using the three measures.
First, know when each measure is appropriate to use. As a rule of thumb, just remember that for an accurate
assessment of return relative to risk, you must consider the actual risk of the investment over the period,
whether systematic only, unsystematic only, or some combination of the two:
The Sharpe ratio measures excess return (above the risk-free rate) per unit of total risk, which is
measured with standard deviation. In this fashion, the analyst makes one of two assumptions:
The portfolio is completely diversified, so the only risk present is systematic risk, or…
The asset is stand-alone, so total risk is the appropriate measure of risk.
The Treynor measure uses beta to measure risk, so it measures return per unit of nondiversifiable
(systematic) risk, only. Treynor is appropriate only when the portfolio is fully diversified (i.e., when
there is no unsystematic risk in the investment).
The CAPM methodology used to calculate Jensen’s Alpha assumes returns of the portfolio are
determined by the market portfolio, and the expected value of the error term is zero. The error term
in the CAPM regression measures unsystematic risk (variability in the portfolio’s returns that is not
28
2004 CFA L 3 – S tudy T ips
explained by the market), so the analyst utilizing Jensen’s Alpha is implicitly assuming only
systematic risk is present. Also, the R2 for the Jensen regression indicates the percentage of
variability in the dependent variable (the portfolio’s excess returns) that is explained by the market’s
excess returns. A very low R2 would indicate significant unsystematic variability (lack of
diversification).
Using Sharpe and Treynor to Rank Portfolios
If a portfolio was not well-diversified over the measurement period, it may be ranked higher using Treynor
than using Sharpe, because Treynor considers only the systematic risk (beta) of the portfolio over the period.
When the Sharpe ratio is calculated for the same portfolio, the increased risk (standard deviation captures
both systematic and unsystematic risk) due to using total risk may cause a lowering in the rankings.
When a portfolio drops in ranking using the Sharpe ratio as compared to the Treynor ratio, it is not as
diversified as the other portfolios, and its Treynor measure, which considers only systematic risk, is not
catching the actual total risk over the measurement period.
If a portfolio rises in ranking using the Sharpe ratio, it was more diversified than the comparison portfolios
(i.e., the comparison portfolios’ standard deviations contained more unsystematic risk, so they dropped in
rank).
The Information Ratio
The information ratio compares the performance of the portfolio to the risk undertaken to achieve the
performance. When portfolio performance is compared to its benchmark, any returns over the benchmark are
considered surplus return. (Note: This is also referred to as tracking error elsewhere in the curriculum.)
The information ratio is the surplus return divided by its standard deviation; therefore, it compares the surplus
generated by the portfolio manager to the risk taken to achieve the surplus:
Again, the information ratio is defined as average tracking error divided by the standard deviation of the
tracking error:
29
2004 CFA L 3 – S tudy T ips
Remember, tracking error is also defined as the standard deviation of the portfolio excess return over the
period, which would be the same as the denominator in the above expression. That is, the difference between
portfolio and benchmark returns is also referred to as excess return (same as the numerator above). Ugh!
This confusion was discussed in my April 22nd tip.
Performance Attribution
Portfolio managers seek to outperform their benchmark by either selecting undervalued securities (selection
effect) or shifting allocation among asset classes at the right time (allocation effect, a.k.a. sector rotation or
market timing). Allocation skill involves shifting money out of sectors that are expected to perform poorly into
those sectors that are expected to perform well. Performance attribution tries to identify the source of a
manager’s performance as superior selection skill, superior allocation skill, or both.
The allocation effect identifies the return attributed to the manager’s decision to change the weights of asset
classes in the portfolio as compared to their weights in the benchmark. In other words, the manager tries to
time the market by weighting heavily in entire sectors he expects to outperform, while lowering the weights of
sectors expected to underperform.
The security selection effect captures the difference in returns caused by the selection of individual assets
within the asset classes. This measure captures the manager’s ability to select superior individual properties,
securities, etc. within the sectors in the portfolio.
Read the equations carefully:
You’ll notice that the first equation, for allocation effect, multiplies the difference in weights (portfolio weight
less benchmark weight) of each asset class (sector) by the excess return earned by the asset class. The
excess return here is measured as the return for the asset class minus the total benchmark return. So, the
equation measures the portion of the return attributable to the manager’s ability to select outperforming asset
classes (sectors).
The second equation, for selection effect, multiplies the weight of the asset class in the portfolio times the
excess return earned by the asset class. The excess return is defined as the return for the asset class in the
30
2004 CFA L 3 – S tudy T ips
portfolio as compared to the same asset class in the benchmark. It measures the manager’s ability to select
superior properties/assets from all those available in the sector.