My goal is to post a simple/short question every day on this thread…so, feel free to answer, and check back daily for others’ answers and feedback. My hope is that over the next 30 days or so, some piece of info on here will stick with you and give you that extra edge on test day (just another avenue of recalling this massive amount of info in the curriculum)… Hopefully these aren’t too easy, or too hard (and short enough so you’re not “wasting time” on here).
Question 1. (4 minutes)
A portfolio manager has been given a mandate of managing a Russell 2000 (small cap) index fund for a large foundation. The manager has the choice between a Full Replication Strategy or Stratified Sampling for managing the portfolio.
Recommend which of the two approaches is preferable. Justify with two reasons.
*note: this is a variation of a BB example in the CFAI text - very testable I think.
Stratified sampling is preferable.
Reasons
(1) The number of stocks are large in russell 2000 which will require a extensive rebalancing.
(2) small cap stocks are less lequid and hence have higher transaction costs than large cap. rebalancing costs are high in such case.
I literally just finished doing that problem in the text, they asked about optimization instead of stratified sampling.
sachin, good work.
spork, hopefully you got this right then 
Hold on, full replication doesn’t need much rebalancing, the stocks will rebalance automatically (assuming market value weighted index). Only rebalance is when new stocks are added and old taken away from the index.
Agree with stratified sampling answer.
and dividends must be reinvested
yes agreed, my both points consider rebalancing which is not correct with respect to full replication.
So the two reasons would be (1) transaction costs due to large number of securities. and (2) higher transaction costs due to limited liquidity of small cap?
if tracking risk is not a concern/requirement - reinforces case for stratified sampling over full replication.
Good points…unless the benchmark type is noted (ie value weighted etc), rebalancing costs only clouds an answer.
Reason 1: when there are more than 1000 securities, full replication should not be used as it is costly and time consuming. Therefore, stratified sampling should be used.
Reason 2: when are a significant of small cap stocks, full replication would be inefficient and trading costs may be high as small stocks tend to be more illiquid. Optimization will enable them to closely match the exposures to the index without having to purchase every security, most likely reducing trading costs.
Question 2:
Using a Completeness Fund would most likely:
a. reduce a portfolio’s Information Ratio
b. eliminate misfit risk
c. Increase the manager’s true active return.
(For added practice, if you’d like, justify your answer)
I think its B because completeness fund try to mimic the benchmark while leave some room for active return. Hence it manages to reduce misfit risk
C
tries to mimic benchmark and also adds room for activemanagement so it will give manager a chance to add true active return.
(guesswork)
do you mind putting in reading number as well? so if we don’t know, we can go back and review easily…
B
A) is the correct answer, methinks
B) a “CF” will reduce misfit risk, but can’t completely eliminate it
C) and by trying to more closely mimic a benchmark, a “CF” will actually decrease a manager’s true active return
B for sure
A is correct. The point of completness fund is to reduce the concentrated position (large active return and risk) by buying other assets that mimic the investors benchmark (not the manager) to have a diversify portfolio. Doing this Will reduces total active return( misfit and true active ) as well as total active risk (misfit + true active risk). B is wrong because it will not eliminate the misfit risk because it still contain assets that were concentrated and following the manager style) c. This is opposite of the purpose of completness fund.
I’m going B
A completeness fund will likely reduce active return, but it will also reduce active risk, so the effect on the information ratio is undetermined. C is not the aim of the completeness fund and the opposite will likely happen.
I don’t like the word ‘eliminate’ but surely a completeness fund can be buiilt with computers to completely mitigate misfit risk, even if it’s only 0 at one point in time. I believe the CFA text states that this is actually a disadvantage of the completeness fund as some misrit risk may be optimal due to tactical allocation.
In other word, if you mange the portfolio actively and your benchmark is A , your true active return must increase as well as truce active risk too. The sponser might have concern about the misfit active return and instruct you to add some assets to your portfolio to mimic the sponser(investor benchmark. Which is Benchamrk B. If you listen to him you will reduces active misfits return and misfit active risk. So if we hold the true active and true risk constant this will reduces both total active return and total active risk. Make sense ?