Shifts in VG Tracking Indexes Announced

[BlueEars]

Here's the CRSP US Methodology Guide
http://www.crsp.com/documentation/pdfs/ ... -Guide.pdf
...

Seems this has disappeared from the site now. I didn't read it yesterday but intended to do it.

[rkhusky]

Do Vanguard and FTSE calculate their median market cap differently? Vanguard lists the median market cap for Total International as $21.5B, with 6286 stocks, whereas the FTSE Global All Cap ex US Index lists its median market cap at $768M, with 5313 stocks.

Vanguard uses the FTSE Global Small Cap Ex Us index for its international small cap. The Vanguard page lists the fund as having 3067 stocks and a median market cap of $1.3B. The FTSE fact sheet lists 3059 stocks with a median market cap of $359M. I thought there would only be one way to calculate the median?!

[ftobin]

Do Vanguard and FTSE calculate their median market cap differently? Vanguard lists the median market cap for Total International as $21.5B, with 6286 stocks, whereas the FTSE Global All Cap ex US Index lists its median market cap at $768M, with 5313 stocks.

Vanguard's number looks more like a mean than median.

[gkaplan]

Gus Sauter is going to tell all in a "live video webcast" on Tuesday.

I assume you mean next Tuesday and not yesterday, the day of the announcement.

[ClosetIndexer]

IMO the mean (average) market caps are the more important number than median anyway. Add a handful of extra small caps and the median will drop very quickly while the mean won't change much at all. It's the mean that will be more closely related to the size factor, and therefore theoretically to returns.

That also seems to make sense since they're looking for lower-cost, but essentially matching indexes. So we would have MSCI EM (old) vs. FTSE EM (new). Aside from the loss of South Korea (and therefore Samsung as the #1 holding), they look basically the same in terms of # of firms and company size. (Note the dates of those fact sheets are off by a month, which is why things like the order of the top holdings is different.)

The MSCI fact sheet above does not seem to match the VG EM fund in terms of median market caps: $2.0B vs. $16.2B (https://personal.vanguard.com/us/funds/ ... st=tab%3A2)

Ah, good point. It looks like the correct MSCI index is the "Select Emerging Markets Index", which was a special Vanguard-only index. Can't find it on MSCI's site, so I guess the Vanguard numbers are probably right, which means the new index should be more diversified into small caps, just like the developed one. Seems tough to gauge just how much more small it is though without either average market cap for the existing funds (is this published somewhere?) or historical returns for the new indexes.

[HueyLD]

Gus Sauter is going to tell all in a "live video webcast" on Tuesday. I could give you the time, but I deleted the email already. (I said could because I know how to dig through my trash)

Register and ask your questions.

Here is the link to register for the webcast http://www-waa-akam.thomson-webcast.net ... dab1313773 .

[davidkw]

Announced on the Nightly Business Report.

HUDSON: Index fund giant Vanguard is changing measuring sticks for 22
of its funds, impacting more than a half-trillion dollars of investor
money. The mutual fund firm says the changes will lead to shareholder
savings. In January, 16 U.S. stock and balanced funds, including this one,
Vanguard’s largest index fund, the total stock market fund, will follow
benchmarks developed by the University of Chicago. Six Vanguard
International stock funds will stop following benchmarks from MSCI
(NYSE:MSCI) and switch to indices operated by the FTSE group based in
London.

Earlier in the program, we reported on how mutual fund company
Vanguard will be changing benchmarks for 22 of its index funds next year.
The business loss for its current provider, MSCI (NYSE:MSCI), is estimated
at $24 million. But that led to more than a $1 billion loss of the stock
market value today for MSCI (NYSE:MSCI), shares plunging 26.8 percent, free
falling to a new low. Analysts note the Vanguard business may be small
financially, but it could change or rather the change could be a knock on
MSCI’s prestige.

Oct 2

[rkhusky]

Do Vanguard and FTSE calculate their median market cap differently? Vanguard lists the median market cap for Total International as $21.5B, with 6286 stocks, whereas the FTSE Global All Cap ex US Index lists its median market cap at $768M, with 5313 stocks.

Vanguard's number looks more like a mean than median.

The FTSE fact sheet lists the average at $3.4B, still a large difference.

[NORDO]

Thanks very much for doing this regression work. Reduces the immediate anxiety for me. Also I really needed that monthly CRSP data link.

+1 to this ... very helpful, thanks!

[grabiner]

For those who enjoy detail, this is a link to the CRSP U.S.Equity Indexes Methodology Guide:

http://www.crsp.com/documentation/pdfs/ ... -Guide.pdf

This document says that Business Development Companies are not included in the CRSP indexes, which will make the expense ratios less confusing.

[pkcrafter]

Does anyone see the irony in Vanguard switching to CRSP indexes?

CRSP being the Center for Research in Security Prices, an extension of the University of Chicago Business School named after David Booth? Booth of course gave away much of his wealth to UoC, which no doubt is used for research efforts, some of which I'm sure have gone into the development of these indexes.

In a way, DFA is helping Vanguard with their indexes, no? YES! And it is, indeed, ironic. I'm feeling a bit uncomfortable with these changes--recharacterization quarterly? Packets of stocks being moved up and down? More complexity? How much DFA slippery stuff will be in these indexes? I wonder....

It's a small world.

[LadyGeek]

Is the CRSP (Center for Research in Security Prices) considered a major index provider? I was wondering if the wiki should be updated.

Wiki article link: Stock Market Indexing

Currently, it only mentions the CRSP as a supplier to the academic community.

This document says that Business Development Companies are not included in the CRSP indexes, which will make the expense ratios less confusing

Wiki article link: Business Development Company

[William4u]

This is what I'd like to see with the old vs new indexes, especially the total domestic and total international...

.................% in old index......... % in new index
1. LG...................??%...........................??%
2. LV...................??%...........................??%
3. MG...................??%...........................??%
4. MV...................??%...........................??%
5. SG...................??%...........................??%
6. SV...................??%...........................??%

So I want to see a comparison between the large mid and small cap percentages of the total indexs for both value and growth. For example, is the new index a little more value, or a little more small cap, or what? I want to see the numbers here.

[ClosetIndexer]

This is what I'd like to see with the old vs new indexes, especially the total domestic and total international...

.................% in old index......... % in new index
1. LG...................??%...........................??%
2. LV...................??%...........................??%
3. MG...................??%...........................??%
4. MV...................??%...........................??%
5. SG...................??%...........................??%
6. SV...................??%...........................??%

So I want to see a comparison between the large mid and small cap percentages of the total indexs for both value and growth. For example, is the new index a little more value, or a little more small cap, or what? I want to see the numbers here.

I posted the regression results for all the domestic indexes above. Factor loadings are more meaningful numbers than "% MG" etc., since that depends entirely on how you define "mid" and "growth". TL;DR: on the domestic side the new indexes are essentially identical to the old from a size and value perspective.

As far as the international indexes, none are style indexes, so we expect their value loadings to be around 0. It looks like the developed and emerging indexes are smaller than the MSCI versions, and the global ex-US all-cap is perhaps slightly larger. Those are just guesses based on average market cap though, since we don't have historical returns for FTSE indexes so can't do regressions.

[dmcmahon]

In a somewhat unrelated news item, Vanguard appears to be adding assets at a torrid rate. If the trend continues for a while, some amount of the needed rejiggering may be possible by taking new money and using it to fill in holes. At least, this may be true for the TSM indexes where there are presumably no big deltas (e.g. +/- an entire country!).

[Barry Barnitz]

Is the CRSP (Center for Research in Security Prices) considered a major index provider? I was wondering if the wiki should be updated.

Wiki article link: Stock Market Indexing

Currently, it only mentions the CRSP as a supplier to the academic community.

This document says that Business Development Companies are not included in the CRSP indexes, which will make the expense ratios less confusing

Wiki article link: Business Development Company

Hi LG,

With the impending move of Vanguard index assets to the new investable CRSP indexes, they will automatically grow in prominence. In light of the Vanguard decision, the Vanguard white paper on benchmarks has been updated (October 2, 2012) to include the CRSP indexes (the table which we reproduce in the wiki page has been updated to include CRSP). See Vanguard - Determining the appropriate benchmark: A review of major market indexes

Investors charged with overseeing a portfolio, either for themselves or another, often evaluate benchmarks to gain perspective on their investments' performance. Absolute results reveal little about why a gain or loss occurred. But relative results, obtained through comparison with market indexes, can shed considerable light on such questions. This paper examines how the major index providers construct and maintain their measures, and offers guidelines for selecting the most appropriate benchmark to evaluate a portfolio's performance.

For the CRSP take on their venture into investable indexes see Chicago Booth Magazine: CRSP Launches Investable Indexes.

regards,

[jon-nyc]

Notice how that news was a blow to MSCI --- their stock is down almost 30% right now.

And it only slightly recovered on Wednesday. Since VG was only ~24MM of revenue out of 900MM+, this is clearly a concern about future pricing power in the industry as opposed to concerns over lost revenue.

[gkaplan]

Vanguard index swap is all about cost, so says Joel Dickson in an interview with Jason Stipp of Morningstart.

http://www.morningstar.com/Cover/videoCenter.aspx?id=569429

[rkhusky]

This is what I'd like to see with the old vs new indexes, especially the total domestic and total international...

.................% in old index......... % in new index
1. LG...................??%...........................??%
2. LV...................??%...........................??%
3. MG...................??%...........................??%
4. MV...................??%...........................??%
5. SG...................??%...........................??%
6. SV...................??%...........................??%

So I want to see a comparison between the large mid and small cap percentages of the total indexs for both value and growth. For example, is the new index a little more value, or a little more small cap, or what? I want to see the numbers here.

I agree. It would be useful to find other mutual funds that track these new indices, so that we can compare the particulars on Morningstar.

[Ketawa]

Unfortunately, I don't think it will be possible to compare the MSCI and CRSP indexes at the same point in time unless someone launches some new funds. I think Vanguard is the only fund provider that currently uses MSCI domestic indexes, and they will be the only ones using the new CRSP indexes.

[rkhusky]

Unfortunately, I don't think it will be possible to compare the MSCI and CRSP indexes at the same point in time unless someone launches some new funds. I think Vanguard is the only fund provider that currently uses MSCI domestic indexes, and they will be the only ones using the new CRSP indexes.

That is unfortunate. How about the international indices? Surely someone must be using FTSE.

[baw703916]

I wonder if part of this has to do with the financial structure. Vanguard is "mutualized" and CRSP is part of an academic institution, so ultimately neither has to try to generate a profit for shareholders. (At least, that's how I think CRSP is structured).

[Barry Barnitz]

Unfortunately, I don't think it will be possible to compare the MSCI and CRSP indexes at the same point in time unless someone launches some new funds. I think Vanguard is the only fund provider that currently uses MSCI domestic indexes, and they will be the only ones using the new CRSP indexes.

That is unfortunate. How about the international indices? Surely someone must be using FTSE.

You can find this information at Search Results for FTSE | ETF Database.

regards,

[KyleAAA]

Does anybody know if the Small Cap Value fund's switch to the CRSP US Small Cap Value Index will make it a deeper value play off the top of their heads? I believe it's somewhere in the range of 2-15% at the bottom of the market cap spectrum but don't know where the current small cap value index falls.

EDIT: Nevermind, I see closetindexer posted the factor weightings above.

[jon-nyc]

Unfortunately, I don't think it will be possible to compare the MSCI and CRSP indexes at the same point in time unless someone launches some new funds. I think Vanguard is the only fund provider that currently uses MSCI domestic indexes, and they will be the only ones using the new CRSP indexes.

MSCI has 1yr, 3yr, 5yr, 10yr performance data on their site, havne't checked CRSP, but they probably do too.

[Ketawa]

MSCI has 1yr, 3yr, 5yr, 10yr performance data on their site, havne't checked CRSP, but they probably do too.

For anyone willing to do a little spreadsheet work, MSCI and CRSP index data is available at the below links. It'll take a little work with the MSCI data, but you can calculate monthly returns and compare whatever domestic indexes you're interested in.

http://www.msci.com/products/indices/country_and_regional/domestic_equity_indices/us/performance.html
http://www.crsp.com/indexes/index-constituent-lists.html

[ClosetIndexer]

That is unfortunate. How about the international indices? Surely someone must be using FTSE.

You can find this information at Search Results for FTSE | ETF Database.

regards,

Unfortunately, as far as I can tell, no one is using these particular FTSE indexes. And FTSE (unlike MSCI and CRSP) does not publish historical index returns. So about all we can do is make guesses based on the fact sheets.

Perhaps if we ask nicely at that Vanguard web conference they can coerce FTSE to release historical returns for these indexes?

[asset_chaos]

The Economist has been quick to publish an article on Vanguard's index switch, online here http://www.economist.com/node/21564217. However, being quick hasn't allowed them to add much that hasn't already been discussed here. They do attribute Gus Sauter confirming the switch saves around 1 basis point of cost:

According to Gus Sauter, Vanguard’s chief investment officer, the switch will enable the company to cut expenses on its funds by around one hundredth of a percentage point.

[LadyGeek]

I updated the wiki: Stock Market Indexing

Changes:
- Updated Figure 2, which now includes FTSE and CRSP
- Inserted CRSP index links and data in a bunch of locations, please double-check the numbers.
- Added FTSE and CRSP to the Index Methodology table

Barry Barnitz's update is here: US index returns - 2012

[jon-nyc]

The Economist has been quick to publish an article on Vanguard's index switch, online here http://www.economist.com/node/21564217. However, being quick hasn't allowed them to add much that hasn't already been discussed here. They do attribute Gus Sauter confirming the switch saves around 1 basis point of cost:

According to Gus Sauter, Vanguard’s chief investment officer, the switch will enable the company to cut expenses on its funds by around one hundredth of a percentage point.

I don't see it. The funds in question have ~550B in them, the total amount paid to MSCI was 24MM a year. Even if the replacements are free, it isn't anywhere near a basis point. (unless they mean a basis point of the dollar amount of expenses, as opposed to a basis point of ER)

[LadyGeek]

I think the following sentence clarifies the use Index-tracking funds: Index fingered:

According to Gus Sauter, Vanguard’s chief investment officer, the switch will enable the company to cut expenses on its funds by around one hundredth of a percentage point. That is a sizeable chunk of Vanguard’s costs: its ETF tracking MSCI’s index of stocks in developed markets outside the United States charges just 0.12% a year.

I interpret that to be a savings of 8.3% (.01% / 0.12%), which is indeed a sizable chunk.

[ClosetIndexer]

With the impending move of Vanguard index assets to the new investable CRSP indexes, they will automatically grow in prominence. In light of the Vanguard decision, the Vanguard white paper on benchmarks has been updated (October 2, 2012) to include the CRSP indexes (the table which we reproduce in the wiki page has been updated to include CRSP). See Vanguard - Determining the appropriate benchmark: A review of major market indexes

For the CRSP take on their venture into investable indexes see Chicago Booth Magazine: CRSP Launches Investable Indexes.

Just wanted to add a +1 for that Vanguard white paper. Very interesting reading. I've looked at it before, but didn't realize it had been updated to include the CRSP indexes.

Reading about CRSP's methodology, along with the fact that the historically back-tested results (from the regressions above) have been so similar to MSCI's suggests to me that this change should be a (very slight) good thing. Which in my mind is ideal, since if the differences were more significant I might think it prudent to switch over to the S&P-based funds (assuming no tax consequences) until these had more of a real-world track record. As it is though, I don't personally see that being necessary.

[ClosetIndexer]

I don't see it. The funds in question have ~550B in them, the total amount paid to MSCI was 24MM a year. Even if the replacements are free, it isn't anywhere near a basis point. (unless they mean a basis point of the dollar amount of expenses, as opposed to a basis point of ER)

That's a good point. Even if he was factoring their planned fee hikes, that doesn't make sense. Perhaps he rounded up to the nearest bp?

Unfortunately you might be right about it meaning "a basis point of the dollar amount of expenses". If the funds have $550B in them and currently charge 0.12%, that's $660M in fees. 24M/660M is about 3.5bps, which gets us in the right ballpark. So if the new indexes aren't free, but just cheaper, it looks about right that the savings could be 0.01% OF the expenses, not 0.01% IN expenses. Boo.

Edit: 24/660 is around 3.5%, not 3.5bps, as pointed out below. So neither explanation appears to make sense.

[Epsilon Delta]

24M/660M is about 3.5bps, which gets us in the right ballpark..

It's a lot closer to 3.5%.

[ClosetIndexer]

24M/660M is about 3.5bps, which gets us in the right ballpark..

It's a lot closer to 3.5%.

Ah, indeed it is. Indeed it is.

OK, so neither possibility makes sense then. Hmm.

[rkhusky]

I think the following sentence clarifies the use Index-tracking funds: Index fingered:

According to Gus Sauter, Vanguard’s chief investment officer, the switch will enable the company to cut expenses on its funds by around one hundredth of a percentage point. That is a sizeable chunk of Vanguard’s costs: its ETF tracking MSCI’s index of stocks in developed markets outside the United States charges just 0.12% a year.

I interpret that to be a savings of 8.3% (.01% / 0.12%), which is indeed a sizable chunk.

Perhaps a sizable chunk of the fee, but not really a significant amount in the grand scheme of things. 0.01% of a $1M is only $100. If that is all that is being saved, I don't know if it is worth the effort, unless the new indices have some significant advantage. I hope Vanguard isn't switching because of some petty tiff between management of the two companies.

[555]

"The Economist has been quick to publish an article on Vanguard's index switch, online here http://www.economist.com/node/21564217. However, being quick hasn't allowed them to add much that hasn't already been discussed here. They do attribute Gus Sauter confirming the switch saves around 1 basis point of cost:"

According to Gus Sauter, Vanguard’s chief investment officer, the switch will enable the company to cut expenses on its funds by around one hundredth of a percentage point.

For Vanguard Institutional Total Stock Market Index Fund Institutional Plus Shares (VITPX)
https://personal.vanguard.com/us/FundsS ... IntExt=INT
that would reduce the expense ratio by 37%. Someone making the minimum investment in that fund would save $20,000/year.

[555]

"Vanguard index swap is all about cost, so says Joel Dickson in an interview with Jason Stipp of Morningstar."
http://www.morningstar.com/Cover/videoCenter.aspx?id=569429

There might also be other reasons that they don't want to say.

[sscritic]

I realize that I am really late to the party, but can someone explain the use of the terms large, mid, and small? If you look at what is listed in Rick Ferri's article, there are mega caps, mid caps, and small caps. Large is the union of mega and mid as I see things and distinct from neither.

Code: Select all

Mega       1 -  300  first 70%
Large      1 -  750  first 85%
Mid      301 -  750  71% - 85%
Small    751 - 2500  85% - 98%

Large, Mid, Small are not disjoint; writing large, mid and small as if they were either makes no sense or is misleading at a minimum. Why don't people write mega, mid, small? I have to admit I have been fooled by this for years, thinking that mid was distinct from large.

Is the same true in the international sphere, does "large" contain every "mid"?

[ClosetIndexer]

I realize that I am really late to the party, but can someone explain the use of the terms large, mid, and small? If you look at what is listed in Rick Ferri's article, there are mega caps, mid caps, and small caps. Large is the union of mega and mid as I see things and distinct from neither.

Code: Select all

Mega       1 -  300  first 70%
Large      1 -  750  first 85%
Mid      301 -  750  71% - 85%
Small    751 - 2500  85% - 98%

Large, Mid, Small are not disjoint; writing large, mid and small as if they were either makes no sense or is misleading at a minimum. Why don't people write mega, mid, small? I have to admit I have been fooled by this for years, thinking that mid was distinct from large.

Is the same true in the international sphere, does "large" contain every "mid"?

Large, mid, small, prime, mega, etc, are defined by the people making the indexes, and each has a different definition. This paper that Barry Barnitz linked to above illustrates each major index provider's definitions very nicely. (Click 'view pdf' on the linked page.)

[vencat]

Can we now expect that Vanguard Small Cap Value (and Large value) will be following the same index as equivalent DFA funds?

[linenfort]

Via twitter, Jason Zweig just shared an article he wrote two years ago introducing the (at the time, new) CRSP indexes: http://online.wsj.com/article/SB1000142 ... hare_tweet

That's a good article, as it makes it completely clear why CRSP is superior (reducing frontrunning, etc.)
I would just like to know, can we pronounce CRSP "CRiSP"?

[Kenster1]

For those who missed the Forbes article by Rick Ferri...

Vanguard Does the CRSP Shuffle

http://www.forbes.com/sites/rickferri/2 ... p-shuffle/

***

Furthermore - it was mentioned in a question here in this thread if any other ETFs track the FTSE international indexes....Yes! Schwab ETFs!

Interestingly, Vanguard's biggest competitor from a low pricing perspective is Schwab and they're already using the FTSE indexes.

Schwab International Largecap ETF (SCHF): 0.09% ER - Tracks FTSE Developed ex-U.S. Index

Schwab Emerging Market ETF (SCHE): 0.15% ER - Tracks FTSE All Emerging Index

Schwab International Small-cap Equity ETF (SCHC): 0.20% ER - Tracks FTSE Developed Small Cap ex U.S. Liquid Index made up of approximately 2,000 international small cap stocks.

[sscritic]

I would just like to know, can we pronounce CRSP "CRiSP"?

Krispy, just like in Kreme.

[BlueEars]

I would just like to know, can we pronounce CRSP "CRiSP"?

Krispy, just like in Kreme.

This is a critically important point . This interview with Gus Sauter has him pronouncing it "crisp": https://personal.vanguard.com/us/insights/video/2426-GusBenchmarks-RET

However, "Krispy" gets the cuteness award.

[ftobin]

However, "Krispy" gets the cuteness award.

I'll make it easier: "Krispy" would be an appropriate pronunciation for the CRSP Indexes: CRSPI

[555]

I wonder if there is a Center for Research in Annuity Prices.

[Barry Barnitz]

Hi:
Note to forum members:

After attempting to compile annual return series from CRSP investable index data, I noted that my return calculations were at variance with the reported returns CRSP made available in their March 2012 and June 2012 factsheets. (See CRSP Index Return Data for my spreadsheet computations). I pm'd ClosetIndexer about this problem and he subsequently e-mailed CRSP about the discrepancy. CRSP responded, and noted that the firm "did not anticipate the very immediate need for data to the extent that we heard it. In responding, we quickly posted returns and constituent lists, and didn’t tie out the results posted in various documents and publications. The result of our good intentions is the discrepancies that you pointed out."

In a later communication: "Again, in our scramble to address demand at a level we didn't anticipate right out of the box, we left some strings untied. We are working diligently to straighten things out and hope to have the monthly returns posted today."

CRSP took down their spreadsheets of backdated return data, and as of 10/08/2012, have yet to publish the corrected return series.

They have, however, posted corrected index constituent listings for the indexes . See CRSP - Index Monthly Returns & Constituent Lists for index holdings as of September 28, 2012.

In addition, CRSP has a convenient listing of news releases concerning the indexes and Vanguard's decision to use them in tracking index funds. See Recent Press Coverage.

[edit: link correction]
regards

[LadyGeek]

CRSP took down their spreadsheets of backdated return data, and as of 10/08/2012, have yet to publish the corrected return series.

They have, however, posted corrected index constituent listings for the indexes . See Index Monthly Returns & Constituent Lists for index holdings as of September 28, 2012.

That link is to the main site, try this instead: Index Monthly Returns & Constituent Lists. So far, all I see are the CRSP Constituents with the corresponding Vanguard fund; nothing on returns. BTW, downloading FULL LIST OF CONSTITUENTS >> gets you all the indexes in one spreadsheet.

Acronym question: What does "SOD" mean? It's the column title in the Constituent list, as "SOD Weight 9/28/2012" with units of %. I think it's percent of total holding.

No comment on CRSP's quality control, I suspect that's being addressed behind the scenes.

[gkaplan]

Acronym question: What does "SOD" mean? It's the column title in the Constituent list, as "SOD Weight 9/28/2012"

http://www.crsp.com/indexes/constituent-lists.html

Index Monthly Returns & Constituent Lists

Choose from the following list of indexes to download Excel workbooks with the associate constituents for each index. Constituents and weights were effective for September 28, 2012. Dates are Start of Day (SOD).