[What is the theory behind "float adjusted" market cap indices?]

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
nirvines
Posts: 27
Joined: Mon Oct 24, 2011 11:09 am

[What is the theory behind "float adjusted" market cap indices?]

Post by nirvines » Fri Oct 28, 2011 9:57 am

We all know the arguments for cap weighting, etc. However, none of the index funds we use truly use a cap weighted approach because they are all "float adjusted" to subtract the holdings of certain share owners such as governments. From a practical perspective, this is understandable because it seems it would be too easy for governments to manipulate shares (for example, by privatizing an agency and then taking a big stake in it they can get dollar for dollar matching from indexers).

But what is the theoretical justification for float adjustment?
1. If government owned shares should be excluded from cap weight, then how do we know that quasai government institutions who have similar power to manipulate and bureaucratic limitations shouldn't also be excluded (indeed in the context of the grant system such institutions are sometimes referred to as "third party government")? For example, Larry Swedroe recently argued that we should avoid corporate bonds due to the bureaucratic limits imposed on institutions...but does this kind of logic not also apply to certain stocks?
2. For that matter, how do we know that we shouldn't also exclude employee holdings, buybacks, and the holdings of other entities who may have distorted incentives and/or a privileged ability to manipulate?
3. Finally, since a large part of this issue seems to boil down to liquidity and whether or not shareholders can or will sell at a fair price, why shouldn't we also exclude the holdings of people like Great Uncle Bob who bought his shares of Con Ed in 1923, has held the paper certificates in his fire safe ever since, and never intends to patron the stock market ever again?

What is the theory of float adjustment that can be used to answer these questions and how does it relate to the theory of cap weighting? Is it possible that strategies like Fama-French tilting and Arnott's Fundamental Indexing are just better ways to adjust the index for float?

User avatar
XdUzHa3NtSeIkBkIGPVn
Posts: 46
Joined: Thu Jul 16, 2015 4:12 pm

Float-adjusted market cap

Post by XdUzHa3NtSeIkBkIGPVn » Mon Nov 16, 2015 4:21 pm

[Thread merged into here, see below. --admin LadyGeek]

What's the rationale for "float adjusting" market capitalization indexes?

I understand that if a 100B$ company only allowed 0.1% of its shares to be publicly traded, there wouldn't be enough shares in the market for (non float-adjusted) indexes to fill their allocations. However, it seems like the line between "restricted shares" and "free floating" is blurry at best.

According to FTSE in its index construction methodology:
Free float is the proportion of shares in issue that are deemed to be tradable.
Their document "Free Float Restrictions" has some guidelines on what counts as restricted, but it only makes the distinction by setting some arbitrary lines:
Free float restrictions include:
[...]
Shares held by Sovereign Wealth Funds where each holding is 10% or greater. If the holding subsequently decreases below 10%, the shares will remain restricted until the holding falls below 7%.
So a SWF can hold 9% of outstanding shares for 10 years, then tick up to 10%. The company's market weight in indexes will then decrease by 10%, and stay down even if the fund goes back to 9% for the next 10 years.

So, this leaves me asking:
  • Is there a theoretical reason for float-adjusting, or is it just done for practical considerations?
  • Are there any data for the performance of float-adjusted and non float-adjusted indexes?

User avatar
nisiprius
Advisory Board
Posts: 38502
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Float-adjusted market cap

Post by nisiprius » Mon Nov 16, 2015 4:43 pm

The theoretical reason involves the theoretical characteristics of the "market portfolio," the securities actually held by the market itself.

If you want (for whatever reason) to mirror the market portfolio, then you only want to mirror the securities that are actually in the market.

I don't know a general answer, but in 2009 Vanguard switched to using the Barclay's U.S. Aggregate Float Adjusted index, while most of its competitors continued to use the Barclay's U.S. Aggregate index, and so far differences have certainly not been large. Blue, VBTLX, Vanguard Total Bond Market Index (float-adjusted); Orange, AGG, iShares Core U.S. Aggregate Bond ETF (not float-adjusted).

Source: Morningstar
Image
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

alex_686
Posts: 4696
Joined: Mon Feb 09, 2015 2:39 pm

Re: Float-adjusted market cap

Post by alex_686 » Mon Nov 16, 2015 5:17 pm

I don’t know of any studies of free-float verse non-free float indexes. However, even if non-free float indexes did better that would not answer your question. Just because a non-free float index would perform better than a free float does not mean an index fund would perform better. The non-free float would have more illiquid stocks and thus higher trading costs. I have heard of bad situations where hot money flows into a small sector.

And a few more points.

When you have a owner with a concentrated position whacky things can happen. A Sovereign Wealth Fund may not pursue profit maximization but some other goal, such as keep jobs within the country. They might be willing to suffer a haircut of 25% to 50% of the value of the company. If you included the SWF shares then you are overweighting the SWF’s judgments over the market’s judgment.

I personally think the more important reason is to handle crossholdings of companies. If company A owns 50% of company B and you did not adjust for market weight one would double count the same shares of B, thus overweighting. Samsung is a great example of this.

User avatar
Lieutenant.Columbo
Posts: 1178
Joined: Sat Sep 05, 2015 9:20 pm
Location: Cicely AK

Re: Float-adjusted market cap

Post by Lieutenant.Columbo » Tue Jun 06, 2017 7:18 am

alex_686 wrote:The non-free float would have more illiquid stocks and thus higher trading costs. I have heard of bad situations where hot money flows into a small sector.
alex_686 wrote:the more important reason is to handle crossholdings of companies. If company A owns 50% of company B and you did not adjust for market weight one would double count the same shares of B, thus overweighting. Samsung is a great example of this.
alex_686, so, you're making the point that investors should follow/use free float-adjusted data, correct?
alex_686 wrote:A Sovereign Wealth Fund may not pursue profit maximization but some other goal, such as keep jobs within the country. They might be willing to suffer a haircut of 25% to 50% of the value of the company. If you included the SWF shares then you are overweighting the SWF’s judgments over the market’s judgment.
if the value of a company goes down, don't the Non-SWF shares go down as well? if so, did Not including SWF shares help the investor that much?
Thanks
Lt. Columbo: Well, what do you know. Here I am talking with some of the smartest people in the world, and I didn't even notice!

User avatar
Lieutenant.Columbo
Posts: 1178
Joined: Sat Sep 05, 2015 9:20 pm
Location: Cicely AK

Re: The theory of float adjustment for indices?

Post by Lieutenant.Columbo » Tue Jun 06, 2017 8:27 am

resuscitating this Topic hoping that those who are familiar with Flote Adjustment care to answer OP's interesting questions
Lt. Columbo: Well, what do you know. Here I am talking with some of the smartest people in the world, and I didn't even notice!

alex_686
Posts: 4696
Joined: Mon Feb 09, 2015 2:39 pm

Re: Float-adjusted market cap

Post by alex_686 » Tue Jun 06, 2017 11:38 am

Lieutenant.Columbo wrote:alex_686, so, you're making the point that investors should follow/use free float-adjusted data, correct?
Mostly yes. A free float market cap index is a very good proxy for that Market Portfolio. In theory the Market Portfolio (a.k.a. all investable assets) is the most efficient portfolio. Now, why would you not want to own, theoretical, the most efficient portfolio? Lots of practical reasons. We could go on a deep dive on some of the nuanced issues. However the consensus is that this should be the neutral starting point for constructing your asset allocation.
Lieutenant.Columbo wrote:
alex_686 wrote:A Sovereign Wealth Fund may not pursue profit maximization but some other goal, such as keep jobs within the country. They might be willing to suffer a haircut of 25% to 50% of the value of the company. If you included the SWF shares then you are overweighting the SWF’s judgments over the market’s judgment.
if the value of a company goes down, don't the Non-SWF shares go down as well? if so, did Not including SWF shares help the investor that much?
Thanks
Yes, if a company's value is impaired it affects everybody, both the SWF and the common investor. BTW, I don't like the term "goes down". These companies tend to increase in value. Its just they offer lower rates of return than a normal company.

Which kinds of brings us back full circle. What are we to do with hybrid companies? Fully including or excluding them is wrong. The easy and internally consistent answer is to use the Market Portfolio, which are investable assets. If large chunks of equity are tied up and are not available to the market, excluded them from the market.

User avatar
David Jay
Posts: 6704
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: The theory of float adjustment for indices?

Post by David Jay » Tue Jun 06, 2017 1:08 pm

Talk about a Zombie thread, the OP posted on the forum for a total of 5 days back in 2011. It is highly unlikely that they are still waiting for an answer.

Why don't you post your own questions with the specific things that are of interest to you?
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius

User avatar
Lieutenant.Columbo
Posts: 1178
Joined: Sat Sep 05, 2015 9:20 pm
Location: Cicely AK

Re: The theory of float adjustment for indices?

Post by Lieutenant.Columbo » Tue Jun 06, 2017 1:36 pm

David Jay wrote:Talk about a Zombie thread, the OP posted on the forum for a total of 5 days back in 2011. It is highly unlikely that they are still waiting for an answer.

Why don't you post your own questions with the specific things that are of interest to you?
I didn't imply that the OP was waiting for an answer, but that I'm interested in learning the answers to those questions. I though it wasn't necessary to start a new Topic only to end up asking the same (or very similar) questions to the ones that had already been asked. All this said, do you have any input on these questions?

User avatar
David Jay
Posts: 6704
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: The theory of float adjustment for indices?

Post by David Jay » Tue Jun 06, 2017 3:48 pm

Angels...
Pin head...
Dancing...
Prediction is very difficult, especially about the future - Niels Bohr | To get the "risk premium", you really do have to take the risk - nisiprius

User avatar
LadyGeek
Site Admin
Posts: 55035
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: [What is the theory behind "float adjusted" market cap indices?]

Post by LadyGeek » Tue Jun 06, 2017 3:59 pm

I merged XdUzHa3NtSeIkBkIGPVn's thread into here, which is a similar discussion.

The wiki has some background info: Stock market indexing (Index methodology)
Wiki wrote:In their attempts to make their indexes more easily investable for index funds, index providers have adopted free-float[16] weighting and buffer zones to their methodologies. Free float weighting eliminates non-trading shares in a company's capital base. These can include cross-ownership of shares by other companies, government owned shares, privately held shares, and other restricted shares. A company's free-float weighting will reflect the actual amount of shares available for public investment.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
Lieutenant.Columbo
Posts: 1178
Joined: Sat Sep 05, 2015 9:20 pm
Location: Cicely AK

Re: Float-adjusted market cap

Post by Lieutenant.Columbo » Tue Jun 06, 2017 9:04 pm

alex_686 wrote:
Lieutenant.Columbo wrote:alex_686, so, you're making the point that investors should follow/use free float-adjusted data, correct?
Mostly yes. A free float market cap index is a very good proxy for that Market Portfolio. In theory the Market Portfolio (a.k.a. all investable assets) is the most efficient portfolio. Now, why would you not want to own, theoretical, the most efficient portfolio? Lots of practical reasons. We could go on a deep dive on some of the nuanced issues. However the consensus is that this should be the neutral starting point for constructing your asset allocation
thank you for your reply;
I take it that, for those wanting to diversify geographically according to the market capitalization, they are better off using the US:ExUS ratio resulting from free float-adjustment compared to using the US:ExUS ratio resulting from Non free float-adjustment, correct?
Last edited by Lieutenant.Columbo on Wed Jun 07, 2017 12:01 pm, edited 1 time in total.
Lt. Columbo: Well, what do you know. Here I am talking with some of the smartest people in the world, and I didn't even notice!

selters
Posts: 646
Joined: Thu Feb 27, 2014 9:26 am

Re: [What is the theory behind "float adjusted" market cap indices?]

Post by selters » Wed Jun 07, 2017 2:00 am

https://www.nbim.no/en/transparency/new ... g-schemes/

https://www.nbim.no/globalassets/docume ... 5ny-14.pdf


Norges Bank Investment Management has published a discussion note on the implications of using free float weighting schemes in global equity portfolios.

The free float of a company is defined as those shares that are readily available for trading, the remaining being held by governments, owners, insiders and other special types of investors. In recent years, free float weighting has largely replaced total market capitalisation as the dominant equity index weighting scheme. The Fund’s strategic equity benchmark, the FTSE Global All Cap Index, is also free float weighted.

In this note, we examine the key implications of using free float-adjustments in global equity portfolios relative to full market capitalisation weights. Both weighting schemes have their pros and cons. Market capitalisation weights can be justified on a theoretical basis, are readily available, easier to calculate, and better represent the relative economic importance of the companies in the portfolio.

The disadvantage of market capitalisation weights is that they do not take into account the trading opportunities in the market and may increase transaction costs. The free float weights mitigate the investability problem but also change the geographic and industry composition of the portfolio and tilt it away from the value, small-cap and less liquid stocks which have been documented to command premia over the long run. This is difficult to justify from a return-maximisation perspective in the case of a long-term investor with no immediate liquidity needs.


Read the whole .pdf for greater detail. 45 pages long.

selters
Posts: 646
Joined: Thu Feb 27, 2014 9:26 am

Re: Float-adjusted market cap

Post by selters » Wed Jun 07, 2017 4:25 am

Lieutenant.Columbo wrote:
alex_686 wrote:The non-free float would have more illiquid stocks and thus higher trading costs. I have heard of bad situations where hot money flows into a small sector.
alex_686 wrote:the more important reason is to handle crossholdings of companies. If company A owns 50% of company B and you did not adjust for market weight one would double count the same shares of B, thus overweighting. Samsung is a great example of this.
alex_686, so, you're making the point that investors should follow/use free float-adjusted data, correct?
I'm sure he is, but you don't really have a choice, because all index funds are free float adjusted.

The non-adjusted index seems to have outperformed if you read the pdf I posted above. But the data provided comes only from one time period. I don't think non-adjusted index data are available anywhere online for free.

It is hard to say which method would be preferable if you could design your own index funds, but to me using non-adjusted weighting (full market cap weighting) seems more logical if you disregard costs, taxes, liquidity and other implementation difficulties, because full market cap weighting better reflects each company's relative economic importance of the in the portfolio. But the index providers and index fund providers seem to have settled for free float adjustment for implementation and cost reasons.

In any case, I don't think this is anything to worry about, because performance is so similar. And only one option is available to us regular investors anyway.

Valuethinker
Posts: 38217
Joined: Fri May 11, 2007 11:07 am

Re: [What is the theory behind "float adjusted" market cap indices?]

Post by Valuethinker » Wed Jun 07, 2017 7:32 am

nirvines wrote:We all know the arguments for cap weighting, etc. However, none of the index funds we use truly use a cap weighted approach because they are all "float adjusted" to subtract the holdings of certain share owners such as governments. From a practical perspective, this is understandable because it seems it would be too easy for governments to manipulate shares (for example, by privatizing an agency and then taking a big stake in it they can get dollar for dollar matching from indexers).

But what is the theoretical justification for float adjustment?
1. If government owned shares should be excluded from cap weight, then how do we know that quasai government institutions who have similar power to manipulate and bureaucratic limitations shouldn't also be excluded (indeed in the context of the grant system such institutions are sometimes referred to as "third party government")? For example, Larry Swedroe recently argued that we should avoid corporate bonds due to the bureaucratic limits imposed on institutions...but does this kind of logic not also apply to certain stocks?
2. For that matter, how do we know that we shouldn't also exclude employee holdings, buybacks, and the holdings of other entities who may have distorted incentives and/or a privileged ability to manipulate?
3. Finally, since a large part of this issue seems to boil down to liquidity and whether or not shareholders can or will sell at a fair price, why shouldn't we also exclude the holdings of people like Great Uncle Bob who bought his shares of Con Ed in 1923, has held the paper certificates in his fire safe ever since, and never intends to patron the stock market ever again?

What is the theory of float adjustment that can be used to answer these questions and how does it relate to the theory of cap weighting? Is it possible that strategies like Fama-French tilting and Arnott's Fundamental Indexing are just better ways to adjust the index for float?
Simply stated, the theory is that you should only benchmark against what you can invest in.

If you will be able to only invest in 5% of Saudi Aramco, post IPO, as seems likely, you can't give it the full $1 trillion value in the index (of its implied market cap).

If you did, every fund manager would try to be market weight, triggering an effective demand for the stock 20x what was actually possible.

When Palm of PalmPilot fame was spun out, there was such an anomaly, as I recall.

Valuethinker
Posts: 38217
Joined: Fri May 11, 2007 11:07 am

Re: [What is the theory behind "float adjusted" market cap indices?]

Post by Valuethinker » Wed Jun 07, 2017 7:39 am

selters wrote:https://www.nbim.no/en/transparency/new ... g-schemes/

https://www.nbim.no/globalassets/docume ... 5ny-14.pdf


Norges Bank Investment Management has published a discussion note on the implications of using free float weighting schemes in global equity portfolios.

The free float of a company is defined as those shares that are readily available for trading, the remaining being held by governments, owners, insiders and other special types of investors. In recent years, free float weighting has largely replaced total market capitalisation as the dominant equity index weighting scheme. The Fund’s strategic equity benchmark, the FTSE Global All Cap Index, is also free float weighted.

In this note, we examine the key implications of using free float-adjustments in global equity portfolios relative to full market capitalisation weights. Both weighting schemes have their pros and cons. Market capitalisation weights can be justified on a theoretical basis, are readily available, easier to calculate, and better represent the relative economic importance of the companies in the portfolio.

The disadvantage of market capitalisation weights is that they do not take into account the trading opportunities in the market and may increase transaction costs. The free float weights mitigate the investability problem but also change the geographic and industry composition of the portfolio and tilt it away from the value, small-cap and less liquid stocks which have been documented to command premia over the long run. This is difficult to justify from a return-maximisation perspective in the case of a long-term investor with no immediate liquidity needs.


Read the whole .pdf for greater detail. 45 pages long.
If you go for a full market cap approach in your investment strategy, then you are implicitly betting that other investors stay with the free float methodology.

Otherwise you are going to see massive scrambles to get stock in a bunch of companies where the Founders (Apple, Amazon, Microsoft) or families still control large chunks. Leading to overvaluation and all kinds of other distortions.

alex_686
Posts: 4696
Joined: Mon Feb 09, 2015 2:39 pm

Re: Float-adjusted market cap

Post by alex_686 » Wed Jun 07, 2017 7:55 am

Lieutenant.Columbo wrote:I take it that, for those wanting to diversify geographically according to the market capitalization, they are better off using the US:ExUS ratio resulting from free float-adjustment compared to using the US:ExUS ratio resulting from from Non free float-adjustment, correct?
I say apples to apples. If you use free float in the US, use free float outside. If market cap in US, then market cap outside.

That being said, for the US market I don't think it matters much if you use market cap or free float. I strongly think free float is the theoretically correct answers but there is not just much difference. Most US companies are pretty "pure" with most of their equity up for trading. Not so outside of the US. There are significantly higher cross holdings and government ownership here, distorting the market.

So use free float for both.

User avatar
Lieutenant.Columbo
Posts: 1178
Joined: Sat Sep 05, 2015 9:20 pm
Location: Cicely AK

Re: [What is the theory behind "float adjusted" market cap indices?]

Post by Lieutenant.Columbo » Wed Jun 07, 2017 8:18 pm

thank you Valuethinker, selters and alex_686 for your replies. I have found them very helpful and easy to understand.
I was surprised at how "readable" selters' Norges .pdf was, and how much of it (I think) I understood.
Lt. Columbo: Well, what do you know. Here I am talking with some of the smartest people in the world, and I didn't even notice!

Post Reply