Buffering Vs Packeting

I do understand the two terminologies however I am confused during practicing to identify the difference.

In CAFI Mock 1, I do not get why did not we pick packeting as the right answer, is it because in case of packeting the transfer would be a portion hence it won’t be refelected as an increase in number of stocks?

Exhibit 2 S&P 500 Index Funds

Manager A** Manager B **Manager C Benchmark S&P 500 S&P 500 S&P 500 Number of holdings: fund/index 498/500 504/500 475/500 Dividends reinvested Next day Same day Next day Management fee (in basis points) 12 15 10 Rebalance Quarterly Quarterly Quarterly Reconstitution Quarterly Quarterly Semi-Annually

Q. As an indexing technique, the number of holdings in Manager B’s index most likely illustrates:

  1. reconstituting.
  2. packeting.
  3. buffering.

I was wondering the same thing. Logically thinking, packeting could increase the number of stocks as well.

i still think packeting was the correct answer.

No, I believe it would be quite the opposite actually. This was a bit tricky of a a question as it requires some understanding of nuance I believe.

Buffering was the right answer because in the end, manager B had 504 stocks (from an index of 500) so he was only holding slightly more than the index is composed of. Buffering is a slow/gradual process of transition so the fact he was only just over was the indication I believe he was doing that.

By comparison, packeting is not gradual, once a stock breaches a the threshold, a certain amount of it is transitioned. How much is moved is subjective but in the end, when you are looking at an entire index of stocks and with all the movement that occurs, they add up and would not be seen as minimal. My guess is if manager B had been packeting, he would’ve held something more like 525-550 or more vs the 504 he had.

But that is my understanding of this, hope it helps. Perhaps if someone more knowledgeable about this aspect can chime in and provide better/further clarification.

This is a good logic but are there any more clarifications on that one?

A buffer places a band around the thresholds that define what goes in an index.

E.g., Large cap may be defined by all stocks >500M in market cap. If a company has a market cap of $505M at the time of reconstitution, it’ll be a large cap stock. If it tanks the next day to $495M, it’s technically not a large cap stock anymore, but if the buffer is set at $485M, it’ll still be a large cap stock. Meanwhile, a stock at $480M market cap may not have been a large cap stock, but due to a good ER report may have gone up to $510M. It’ll still be considered a mid-cap stock if the upper buffer for large cap is set at $515M. As a PM that follows this index, you may have to choose a few stocks in addition to those in the index that are on the boundary to reduce tracking error, and also to reduce transaction costs and bid-ask spreads when trying to rebalance to match the benchmark.

I still don’t understand why packeting wouldn’t also be correct. Anyone have any further thoughts?

Elsewhere in the text, it says that buffering makes the index more investable compared to packeting. Based on the apparent preference the CFAI has about packeting, I will lean that way first on any related exam question.

This is helpful. Thanks a lot. I appreciate it

Concerning packeting, can someone please explain what splitting stock positions into multiple parts mean?

Does it mean that if a manager with a small cap and mid cap mandates owns stock A in the small cap fund, and its market cap increases over the threshold to qualify as a mid cap, he won’t transfer the whole position in the mid-cap fund but just a portion of it? Holding this stock in both the small and mid cap fund?

My take on the whole subject of Buffering vs. Packeting:

  1. A buffer would mean breathing the threshhold by not margin but by miles. Just to ensure that it is a not a momentary blip which the Managers would regret afterwards. So for instsnce if a stock migrates from small cap to mid cap it should just not crawl into mid cap (band 11) but do so in a resounding fashion. This means a small % of such stocks should at least reach the boundary just shy of the Large cap. This gives enough confidence and breathing space for the manager to ascertain that the stock indeed is a mid Cap category stock

  2. Increase of packeting a sizeable portion must have reached that boundary and stay put there for considerable period. This validates the conviction the stock manager places on the buffer indeed the number of stocks in case of buffering would be marginally higher then that of the index where as the number of stocks in case of packeting will be sizeably higher than that of the index. In our example which means 505/500 does qualify for buffering where as 525 / 500 would have qualified for packeting had that one of the choices.

1 Like

just a small clarification requested:

  1. In buffering the stocks which continue to lie in the band (lets say +/-10%) will continue to stay there until it crosses the threshold. Let’s say a mid cap stock crossed the lower threshold for large cap stock, it will continue to be considered as mid-cap until it crosses the actual cutoff and similarly a large cap stock breaching the exact mcap limit of large cap, until the lower limit of mcap is breached. Hence at one time there is a possibility that at one time more than the actual stocks kept in index are present in the replicated portfolio.
  2. In Packeting: when is the it decided a stock will be completely removed from the index. And how is the Percentage of a stock being portion of various classifications being considered (viz: large, mid cap) decided.

Hey Caroline and everyone, how you guys doing?

My understanding is that the logic behind the question has to do with the quantity of stocks you have on the index vs. quantity of stocks you have on the fund. Reconstitution dates discrepancies between the two may lead to the correct answer.

  1. Packeting: When fund’s holdings are <= holdings in the index, it’s probably the case of packeting, because when the established threshold is reached, the holding will appear both on the index you are using as benchmark and on the new index (this will be true until next index reconstitution date). So you end up with more holdings in the index than after the previous reconstitution (think of a multiplicative process, theoretically you could have all the holdings in more than an index at the same time). On the other hand, the fund should wait until its next reconstitution date to get rid or to add a new holding. It means the fund will always carry less (or equal) quantity of holding than the index (because the latter can only increase the number of holdings);

  2. Buffering: When fund’s holdings are very similar to index’ holdings. I would say that buffering tends to have a very similar (and often equal) quantity of holdings on index and on the fund, due to the very orderly and gradual changes that may occur on the index (I mean, the adjust should be very quick and effective, would say even predictable). Now, if you asked me what would be the question’s answer if the Manager B had 500/504 instead of 504/500, we could never be sure which was the correct answer.

Not sure if I helped. But this is a very tricky one…


where did you get CFAI mock? it seems they don’t have it anymore.

This reasoning makes sense.