Are Large-Cap Stock Indexes Always Growth-Tilted?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
SimpleGift
Posts: 3000
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by SimpleGift » Wed Oct 10, 2012 5:26 pm

A fellow "chart enthusiast" recently sent me the chart below, showing the changing sector composition of the S&P 500 over the last 38 years, from 1974-2012. It immediately challenged one of my unconscious, long-held assumptions: that cap-weighted, large company indexes, by the nature of their construction, are always over-allocated to growth stocks.

Image

Certainly, growth stocks dominated the large-cap indexes in the late 1990s during the tech boom. But during the energy bubble of the late 1970s-early 1980s and the financial bubble of the mid-2000s, it appears these indexes were dominated by traditional value sectors. The relative performance of the large-cap growth and value indexes since 1992 seems to confirm this:

(Blue = Growth Index; Orange = Value Index)
Image

Question: Are large-cap stock indexes sometimes growth-tilted and at other times value-tilted? Or am I missing something here? What would a Fama-French value-factor analysis show over time?
Cordially, Todd

User avatar
Archie Sinclair
Posts: 409
Joined: Sun Mar 06, 2011 2:03 am

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Archie Sinclair » Wed Oct 10, 2012 6:05 pm

My understanding is that "Value" and "Growth" are usually defined as one half of the total market each. If every "Value" stock were to disappear tomorrow, we would just reclassify half the Growth stocks as Value stocks. So by definition, neither can dominate the market.

User avatar
baw703916
Posts: 6673
Joined: Sun Apr 01, 2007 1:10 pm
Location: Seattle

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by baw703916 » Wed Oct 10, 2012 7:37 pm

Relative to book value, they are always growth-tilted.

Relative to market cap, they are never growth-tilted nor value-tilted.

It depends on how you want to look at things...
Most of my posts assume no behavioral errors.

User avatar
SimpleGift
Posts: 3000
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by SimpleGift » Wed Oct 10, 2012 8:04 pm

baw703916 wrote:Relative to book value, they are always growth-tilted.

Relative to market cap, they are never growth-tilted nor value-tilted.

It depends on how you want to look at things...
This, perhaps, is the source of my confusion. Often on the Forum, one will hear value-tilting advocates refer to Vanguard's Total Stock Market Fund as "just a large-cap growth fund." I can see how technology companies in the late 1990s had overwhelming market value compared to their book value, and how the index would thus be very growth-tilted. But energy companies, which dominated the large-cap indexes in the late 1970s-early 1980s, would appear to have had a large book value to go with their market value — which would seem to still qualify them as value stocks. Probably I'll need to find a study that does a mechanical, value-factor analysis (book value to market value, or some other value measure) for the S&P 500 over an extended time period. Thanks!
Cordially, Todd

hafius500
Posts: 217
Joined: Sat Sep 10, 2011 4:31 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by hafius500 » Thu Oct 11, 2012 5:33 am

Simplegift wrote:
baw703916 wrote:Relative to book value, they are always growth-tilted.

Relative to market cap, they are never growth-tilted nor value-tilted.

It depends on how you want to look at things...
This, perhaps, is the source of my confusion. Often on the Forum, one will hear value-tilting advocates refer to Vanguard's Total Stock Market Fund as "just a large-cap growth fund." I can see how technology companies in the late 1990s had overwhelming market value compared to their book value, and how the index would thus be very growth-tilted. But energy companies, which dominated the large-cap indexes in the late 1970s-early 1980s, would appear to have had a large book value to go with their market value — which would seem to still qualify them as value stocks. Probably I'll need to find a study that does a mechanical, value-factor analysis (book value to market value, or some other value measure) for the S&P 500 over an extended time period. Thanks!
I've never seen a theoretical and empirical proof of such a permanent growth-tilt. But if an assertion is repeated very often, people will begin to believe it is true.
But here are some regressions:

From Kenneth Scislaw's paper (link here. page 106 pdf:
Results in this essay regressing returns of various Wilshire, S&P, and Russell index series on the Fama and French 3-factor model show that most indexes exhibit a strong value tilt in returns, not a growth tilt. Indeed, the S&P 500 index is the only series to reflect a neutral stance between the two styles...
[page 117] Curiously, the Wilshire Target Small growth series (WTSg), the small cap S&P 600/Citigroup pure growth (sp600pg) shown in Panel A (4), and the broader S&P 1500/Citigroup pure growth series (sp1500pg) shown in Panel A (2) each load positively on the HML factor. This result indicates a value orientation in returns despite a growth construction.
I recall other regressions (by Robert T?) showing a value tilt for the SP500.
BTW, price-to-book ratios do not necessarily predict growth/value-tilts.
A growth-tilted portfolio correlates with the risk factor. That's all.
Assume the value risk is related to recessions. If all of the large companies had such recession risks and if the small companies had no recesssion risks, would the market portfolio be growth-tilted despite the large-caps having lower price-to-book ratios?
prior username: hafis50

richard
Posts: 7961
Joined: Tue Feb 20, 2007 3:38 pm
Contact:

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by richard » Thu Oct 11, 2012 5:46 am

Whether an index is tilted depends on how you define value and growth.

Using the Fama French definitions, US total stock market is not tilted. In other words, the factor loadings for SMB (size tilt) and HML (value tilt) are zero when you run the FF three factor model regression on TSM.

larryswedroe
Posts: 15625
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by larryswedroe » Thu Oct 11, 2012 7:49 am

First the market is not 50/50 split, its top 30% is growth, bottom 30% value and middle 50% is core, at least that is the way academics view it. Which btw is how you get differences in index funds when some indices split their index in half between value and growth. Which is why it matters what the index you benchmark against is.

Second, the market does drift

I happen to have kept some records I dug up 12/99 all percentages rounded

LG 69%
LC 20%
LV 6%
SG,SC, and SV all 2% or less
Small was 5.3% and value 7.5%

12/08
LG 74%
LC 16%
LV 4%
SG 2%
SC 2%
SV 1%

NewtonsApple
Posts: 73
Joined: Thu Feb 24, 2011 10:07 am

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by NewtonsApple » Thu Oct 11, 2012 8:28 am

larryswedroe wrote:First the market is not 50/50 split, its top 30% is growth, bottom 30% value and middle 50% is core, at least that is the way academics view it. Which btw is how you get differences in index funds when some indices split their index in half between value and growth. Which is why it matters what the index you benchmark against is.
I 110% agree. :D

User avatar
baw703916
Posts: 6673
Joined: Sun Apr 01, 2007 1:10 pm
Location: Seattle

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by baw703916 » Thu Oct 11, 2012 8:49 am

hafius500 wrote:

I've never seen a theoretical and empirical proof of such a permanent growth-tilt. But if an assertion is repeated very often, people will begin to believe it is true.
But here are some regressions:

From Kenneth Scislaw's paper (link here. page 106 pdf:
Results in this essay regressing returns of various Wilshire, S&P, and Russell index series on the Fama and French 3-factor model show that most indexes exhibit a strong value tilt in returns, not a growth tilt. Indeed, the S&P 500 index is the only series to reflect a neutral stance between the two styles...
[page 117] Curiously, the Wilshire Target Small growth series (WTSg), the small cap S&P 600/Citigroup pure growth (sp600pg) shown in Panel A (4), and the broader S&P 1500/Citigroup pure growth series (sp1500pg) shown in Panel A (2) each load positively on the HML factor. This result indicates a value orientation in returns despite a growth construction.
I recall other regressions (by Robert T?) showing a value tilt for the SP500.
BTW, price-to-book ratios do not necessarily predict growth/value-tilts.
A growth-tilted portfolio correlates with the risk factor. That's all.
Assume the value risk is related to recessions. If all of the large companies had such recession risks and if the small companies had no recesssion risks, would the market portfolio be growth-tilted despite the large-caps having lower price-to-book ratios?
Really, my point was that it matters how you define growth/value, and what you measure a fund's tilt relative to. If you define growth as the 30% of the total market cap with the lowest BtM and value as the highest 30% by BtM, then a total market fund weighted by market cap will always hold 30% value and 30% growth, simply by mathematical identity. If you define growth as the 30% of total market book value with the lowest BtM, then a market cap weighted fund will always overweight growth, because growth companies are defined in terms of having higher market caps relative to book value.
Most of my posts assume no behavioral errors.

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Thu Oct 11, 2012 11:10 am

Simplegift,

When I use the Ken French Large Growth/Value indexes on his website along with the S&P 500, I find that the large cap index (S&P 500) sits much closer to growth indexes than value indexes. Here are the specific HmL coefficients since 1926:

FF US Large Growth Index = -0.23
S&p 500 = +0.05
FF US Large Value Index = +0.82

So, you see, there is about 0.28 difference between LG and the S&P 500, and 0.77 difference between S&P 500 and LV. Meaning that the S&P 500 is really sort of a watered down growth index (due to its cap weighting and the fact that companies get to be the biggest because of rising share prices).

Incidentally, the results don't change if we use a large cap index other than the S&P 500. The CRSP 1-5 Index (similar to the Russell 1000 or MSCI 750) has a HmL coeff of -0.01, and the CRSP 1-10 Index (total market) comes in at a HmL coeff of 0.00.

In order to find that large cap indexes are neutral between growth and value, we would need to see a HmL coeff of around +0.3 or so (the midway point between -0.23 and +0.82).

That is why you will often see the S&P 500 or TSM substituted for a pure LG index in a balanced asset class portfolio -- slightly higher expected returns while still participating in growth-led markets.

This fact also throws cold water on the notion that a TSM portfolio is well diversified across all stocks. IT HOLDS all stocks, but is heavily concentrated in the largest, most growth oriented businesses. Thats fine as long as you know it, but its important that you do. The "total" in Total Stock Market is a bit misleading and a marketing contract of Wall Street IMO.
The most disciplined investor in the world.

User avatar
SimpleGift
Posts: 3000
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by SimpleGift » Thu Oct 11, 2012 11:56 am

Thanks, Jerry_lee, that's the type of value-factor analysis I was looking for.

However, back to my original question in the OP: Does the tilt to growth or value in the large cap indexes change over time? If so, wouldn't the value-factor loading depend upon the time period that one examined? Following up on some studies suggested by hafius500 above, I found these results for the S&P 500, over these specific time periods:

Looking at the S&P 500 from 1995-2006:
Three-factor model regression results in Table 4 suggest that the S&P 500 index (sp500) does not exhibit a perceptible tilt toward growth or value. The HML coefficient for the index shown in Panel A (1) is 0.02 and statistically indistinguishable from zero (t = 1.14).
K.E. Scislaw (2010), Page 114, Three Essays on the Value Premium: Can Investors Capture the Promised Rewards?

Looking at the S&P 500 from 1975-2002:
Houge and Loughran (2006) also fail to observe a statistical tilt in the S&P 500 index when excess monthly index returns are regressed on the Carhart 4-factor (momentum) model. Also in that study, regression coefficients for the now defunct S&P 500/BARRA value and growth style index series, representing two equal halves by market cap of the S&P 500 Index, load virtually identically (but opposite sign) on the HML book-to-market factor (value = 0.32, growth = -0.30).
Houge and Loughran (2006), Do Investors Capture the Value Premium?

Personally, I'm agnostic on the whole value-factor debate. I'm just trying to understand the behavior of the large-cap indexes over time, and whether they are always growth-tilted or just sometimes.
Last edited by SimpleGift on Thu Oct 11, 2012 12:17 pm, edited 1 time in total.
Cordially, Todd

hafius500
Posts: 217
Joined: Sat Sep 10, 2011 4:31 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by hafius500 » Thu Oct 11, 2012 12:10 pm

baw703916 wrote:
hafius500 wrote:

I've never seen a theoretical and empirical proof of such a permanent growth-tilt. But if an assertion is repeated very often, people will begin to believe it is true.
But here are some regressions:

From Kenneth Scislaw's paper (link here. page 106 pdf:
Results in this essay regressing returns of various Wilshire, S&P, and Russell index series on the Fama and French 3-factor model show that most indexes exhibit a strong value tilt in returns, not a growth tilt. Indeed, the S&P 500 index is the only series to reflect a neutral stance between the two styles...
[page 117] Curiously, the Wilshire Target Small growth series (WTSg), the small cap S&P 600/Citigroup pure growth (sp600pg) shown in Panel A (4), and the broader S&P 1500/Citigroup pure growth series (sp1500pg) shown in Panel A (2) each load positively on the HML factor. This result indicates a value orientation in returns despite a growth construction.
I recall other regressions (by Robert T?) showing a value tilt for the SP500.
BTW, price-to-book ratios do not necessarily predict growth/value-tilts.
A growth-tilted portfolio correlates with the risk factor. That's all.
Assume the value risk is related to recessions. If all of the large companies had such recession risks and if the small companies had no recesssion risks, would the market portfolio be growth-tilted despite the large-caps having lower price-to-book ratios?
Really, my point was that it matters how you define growth/value, and what you measure a fund's tilt relative to. If you define growth as the 30% of the total market cap with the lowest BtM and value as the highest 30% by BtM, then a total market fund weighted by market cap will always hold 30% value and 30% growth, simply by mathematical identity. If you define growth as the 30% of total market book value with the lowest BtM, then a market cap weighted fund will always overweight growth, because growth companies are defined in terms of having higher market caps relative to book value.
I'm not sure I understand you:
If you define growth as the 30% of the total market cap with the lowest BtM and value as the highest 30% by BtM, then a total market fund weighted by market cap will always hold 30% value and 30% growth, simply by mathematical identity
I suspect this is not necessarily true if the companies are sorted by numbers instead of market capitalisation.
If you define growth as the 30% of total market book value with the lowest BtM, then a market cap weighted fund will always overweight growth, because growth companies are defined in terms of having higher market caps relative to book value
Who uses this definition?

I think the difference between the two definitions is that the second definiion denotes a trading strategy (an example of fundamental indexing).
Using the first definition you can recommend that the investor X should own all growth stocks and Y should all value and neutral stocks. The sum of all portfolios would be the market portfolio. Furthermore, everybody can hold the market portfolio.

But everybody cannot hold a cap-weighted portfolio of the total book market value. And you cannot recommend that X should own the top 50% of that portfolio and Y should own the bottom 50%.

If this is true, such a definition of growth/value would not make sense. Or can we use such a definition to formulate a rational market model?
prior username: hafis50

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Thu Oct 11, 2012 12:30 pm

Simplegift wrote:Thanks, Jerry_lee, that's the type of value-factor analysis I was looking for.

However, back to my original question in the OP: Does the tilt to growth or value in the large cap indexes change over time? If so, wouldn't the value-factor loading depend upon the time period that one examined? Following up on some studies suggested by hafius500 above, I found these results for the S&P 500, over these specific time periods:

Looking at the S&P 500 from 1995-2006:
Three-factor model regression results in Table 4 suggest that the S&P 500 index (sp500) does not exhibit a perceptible tilt toward growth or value. The HML coefficient for the index shown in Panel A (1) is 0.02 and statistically indistinguishable from zero (t = 1.14).
K.E. Scislaw (2010), Page 114, Three Essays on the Value Premium: Can Investors Capture the Promised Rewards?

Looking at the S&P 500 from 1975-2002:
Houge and Loughran (2006) also fail to observe a statistical tilt in the S&P 500 index when excess monthly index returns are regressed on the Carhart 4-factor (momentum) model. Also in that study, regression coefficients for the now defunct S&P 500/BARRA value and growth style index series, representing two equal halves by market cap of the S&P 500 Index, load virtually identically (but opposite sign) on the HML book-to-market factor (value = 0.32, growth = -0.30).
Houge and Loughran (2006), Do Investors Capture the Value Premium?

Personally, I'm agnostic on the whole value-factor debate. I'm just trying to understand the behavior of the large-cap indexes over time, and whether they are always growth-tilted or just sometimes.
You are welcome. To answer your question: not really, the S&P 500 always leans more towards growth than value. There are of course some periods where the lean isn't quite as great as others, but it's still. To see this, lets break the last 4 decades in quarters:

1973-1982
LG HmL = -0.33
S&P 500 HmL = 0.00
LV HmL = +0.72

1983-1992
LG HmL = -0.44
S&P 500 HmL = 0.02 <-------- I guess this as close as it gets, but the ann returns for this period were +15.9% LG, +16.2% S&P, and +17.9% LV, clearly a growth tilt
LV HmL = +0.62

1993-2002
LG HmL = -0.27
S&P 500 HmL = 0.01
LV HmL = +0.79

2003-8/2012
LG HmL = -0.29
S&P 500 HmL = 0.03
LV HmL = +0.73

What about the papers you mention? Well, just going off of your quotes, there are a few issues. In the first, the author takes a 0.0 HmL tilt on the S&P 500 to imply no growth tilt. But to properly measure this (remember, we are asking if cap weighted indexes are more "growth-like", or more "value-like", you need to compare it's HmL coefficient to that of a distinct LG and LV indexes to see it's relative orientation). In doing so, it is clear that it leans more towards growth.

In the second, the author is just using circular logic. If you split anything 50/50, where growth = the most expensive 50% of stocks, and value = all stocks not in the most expensive 50%, then by rule, all the stocks (the 50% growth and 50% value) mustl be smack dab in between the two. This approach to indexconstruction forces this to be the case.

Now, if you instead look at growth/neutral/value distinctly, or at least growth/value distinctly, with "neutral" being all the stocks, you can measure how the market orients with the purest attributes of each end of the price dimension. In doing so, you see that a cap weighted large cap index always leans towards the highest priced stocks.

This is one of these things you don't really need to think a lot about. Because when you cap weight all companies, you of course are going to have a portfolio that has the higest proportion in the highest priced stocks. And high prices = growth companies. Low prices = value companies. Of course we need to scale that by a fundamental metric (as in Price TO Earnings) to more precisely define growth/value, but even without doing so, you get to about the same place. Price is what matters, the fundamental metric is secondary.
The most disciplined investor in the world.

User avatar
SimpleGift
Posts: 3000
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by SimpleGift » Thu Oct 11, 2012 12:45 pm

Thanks again for your follow-up, Jerry_lee.

What I take away from your recent post is that, even in periods when the large-cap indexes are dominated by traditional value sectors (energy stocks in the late 1970s-early 1980s; financials in the mid-2000s) — and the large-cap value indexes are strongly outperforming the large-cap growth indexes — somehow the overall market index is still decidedly growth-tilted. At least that's what your factor analysis appears to show, for example, during the 1973-1982 decade.

I'll need to think on this a bit more and understand exactly why this is so. Cheers!
Cordially, Todd

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Thu Oct 11, 2012 12:50 pm

Simplegift wrote:Thanks again for your follow-up, Jerry_lee.

What I take away from your recent post is that, even in periods when the large-cap indexes are dominated by traditional value sectors (energy stocks in the late 1970s-early 1980s; financials in the mid-2000s) — and the large-cap value indexes are strongly outperforming the large-cap growth indexes — somehow the overall market index is still growth-tilted. At least that's what your factor analysis appears to show, for example, the 1973-1982 decade.

I'll need to think on this a bit more and understand exactly why this is so. Cheers!
Yes. The issue you are having is confining a sector to a "value bucket" or a "growth bucket". Obviously, if an industry has high earnings, and rapidly rising prices, that is a growth industry regardless of their products/services. These things change over time. The automotive industry used to be a growth industry, today, well...Telecom used to be a growth industry, today I see a lot of these companies in value baskets.

Old-school Wall Street nominclature is what creates this view that industries are static entities that always fit in one bucket or another, but the market isn't static, its fluid. Think creative destruction.
The most disciplined investor in the world.

rkhusky
Posts: 5200
Joined: Thu Aug 18, 2011 8:09 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by rkhusky » Thu Oct 11, 2012 12:57 pm

[quote="Jerry_lee"
This fact also throws cold water on the notion that a TSM portfolio is well diversified across all stocks. IT HOLDS all stocks, but is heavily concentrated in the largest, most growth oriented businesses. Thats fine as long as you know it, but its important that you do. The "total" in Total Stock Market is a bit misleading and a marketing contract of Wall Street IMO.[/quote]

But isn't the reason that those large companies dominate the TSM is because thousands of traders, poring over thousands of pages of business information, set the market price of those companies based on all available information. The traders must feel that those large companies offer the best risk-adjusted future return of their investment dollars. So, by investing in a manner contrary to the market weight, one is making a bet that those thousands of traders are wrong. That seems to me to be true even if the traders are not investing for long term profitability, but only for short term returns. It is the only measure we have for the value of a stock.

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Thu Oct 11, 2012 1:24 pm

rkhusky wrote:
Jerry_lee wrote: This fact also throws cold water on the notion that a TSM portfolio is well diversified across all stocks. IT HOLDS all stocks, but is heavily concentrated in the largest, most growth oriented businesses. Thats fine as long as you know it, but its important that you do. The "total" in Total Stock Market is a bit misleading and a marketing contract of Wall Street IMO.
But isn't the reason that those large companies dominate the TSM is because thousands of traders, poring over thousands of pages of business information, set the market price of those companies based on all available information. The traders must feel that those large companies offer the best risk-adjusted future return of their investment dollars. So, by investing in a manner contrary to the market weight, one is making a bet that those thousands of traders are wrong. That seems to me to be true even if the traders are not investing for long term profitability, but only for short term returns. It is the only measure we have for the value of a stock.
No, that isn't why. Big stocks are big and prices are high because the companies are very big and very successful. They don't have the best risk-adjusted returns, just the lowest risk (and lowest expected returns). And investing not according to market weight just means you don't want the risk/return profile of the market portfolio. For all its diversification (# of stocks), a few dozen still make up a large part of the whole. If that isn't what you want your portfolio to look like, then hold other parts of the market in a larger proportion. So long as their expected returns are a function of a risk/return relationship, then you are just altering your overall portfolio profile.

If the world market cap portfolio is 60% equity and 40% fixed income, am I "betting" by holding more or less bonds even if that is appropriate for my situation? The sin isn't in deviating from the market, it is doing so in a concentrated way (holding just a handful of securities) that increases unsystematic/ "bad risks" -- the type we aren't rewarded for taking.

The market portfolio is just another arbitrary allocation that can be attatained with really low costs and minimal tax consequences. Don't get me wrong, I love it as a larger/growth holding. I just see many reasons not to be 100% LG.
The most disciplined investor in the world.

User avatar
camontgo
Posts: 324
Joined: Tue Oct 27, 2009 1:46 pm
Location: Chicago, IL
Contact:

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by camontgo » Thu Oct 11, 2012 1:25 pm

Simplegift wrote:However, back to my original question in the OP: Does the tilt to growth or value in the large cap indexes change over time? If so, wouldn't the value-factor loading depend upon the time period that one examined? Following up on some studies suggested by hafius500 above, I found these results for the S&P 500, over these specific time periods:
I posted a graphic of the percent of TSM market cap which fell into each of Fama and French's six buckets (the ones they use to compute the HML and SMB factors) in 2011 here: http://www.bogleheads.org/forum/viewtop ... 5#p1508795

These percentages change over time. Here is the year by year data. Max in LG bucket was 79.1% in 2000, min in LG bucket was 30.9% in 1984. This data could be used to construct a chart similar to that in the OP.

Code: Select all

       SG     SB     SV     LG      LB      LV
1926	1.4%	2.1%	1.7%	37.5%	51.4%	5.9%
1927	1.2%	1.9%	1.6%	49.8%	37.7%	7.8%
1928	0.9%	1.8%	1.8%	45.4%	44.5%	5.5%
1929	0.9%	1.6%	1.4%	49.8%	39.3%	7.0%
1930	0.7%	1.4%	1.1%	64.6%	28.5%	3.8%
1931	0.4%	1.3%	1.0%	70.6%	24.9%	1.8%
1932	0.3%	1.1%	0.8%	76.2%	18.8%	2.9%
1933	0.4%	1.6%	1.2%	68.9%	21.6%	6.3%
1934	0.5%	1.4%	1.1%	52.5%	38.5%	6.0%
1935	0.4%	1.2%	0.9%	50.0%	41.8%	5.7%
1936	0.7%	1.6%	1.1%	51.4%	35.8%	9.3%
1937	1.1%	1.6%	1.2%	43.1%	40.9%	12.1%
1938	0.8%	1.6%	1.2%	62.2%	27.5%	6.7%
1939	0.8%	1.4%	1.1%	51.3%	40.4%	5.0%
1940	0.7%	1.7%	1.2%	60.5%	30.5%	5.5%
1941	0.8%	1.7%	1.4%	59.5%	28.5%	8.1%
1942	0.7%	1.8%	1.4%	63.4%	27.7%	5.1%
1943	1.0%	2.2%	1.7%	62.3%	23.6%	9.2%
1944	1.3%	2.5%	1.8%	55.2%	29.9%	9.4%
1945	1.7%	2.8%	2.0%	42.8%	35.8%	14.9%
1946	2.0%	3.3%	2.6%	37.7%	35.9%	18.5%
1947	1.5%	3.0%	2.2%	40.7%	39.3%	13.2%
1948	1.7%	2.9%	2.1%	43.1%	35.6%	14.6%
1949	1.2%	2.5%	2.1%	58.5%	26.9%	8.7%
1950	1.2%	2.5%	2.2%	50.4%	34.3%	9.4%
1951	1.0%	2.6%	2.3%	50.9%	33.5%	9.7%
1952	1.1%	2.5%	1.9%	55.7%	29.8%	8.9%
1953	0.9%	2.6%	2.0%	56.5%	29.9%	8.1%
1954	0.8%	2.3%	1.9%	60.2%	28.3%	6.6%
1955	0.8%	2.3%	1.9%	60.1%	28.5%	6.3%
1956	0.8%	2.1%	1.7%	59.4%	30.3%	5.6%
1957	0.7%	2.0%	1.7%	64.2%	26.3%	5.0%
1958	0.6%	2.2%	1.7%	62.5%	28.1%	4.8%
1959	0.9%	2.2%	1.8%	60.0%	29.0%	6.2%
1960	0.9%	2.0%	1.7%	54.8%	34.0%	6.6%
1961	1.0%	2.2%	1.8%	62.3%	25.8%	7.0%
1962	1.0%	2.0%	1.8%	49.0%	38.9%	7.3%
1963	1.4%	2.4%	2.0%	60.7%	25.9%	7.6%
1964	1.4%	2.1%	1.9%	58.5%	28.6%	7.5%
1965	1.3%	2.1%	2.1%	55.3%	33.0%	6.1%
1966	1.8%	2.7%	2.2%	50.5%	34.3%	8.6%
1967	1.9%	2.9%	2.6%	51.7%	34.0%	6.8%
1968	3.7%	2.9%	2.8%	43.7%	29.8%	17.0%
1969	3.9%	3.0%	2.4%	38.1%	31.6%	21.0%
1970	2.9%	2.7%	2.1%	40.3%	38.3%	13.7%
1971	2.5%	3.3%	2.7%	50.6%	33.2%	7.8%
1972	2.5%	3.2%	2.5%	54.0%	24.2%	13.7%
1973	2.8%	2.7%	2.2%	54.7%	29.4%	8.2%
1974	2.5%	3.2%	2.2%	62.7%	24.4%	5.0%
1975	2.8%	3.7%	2.5%	65.9%	21.9%	3.3%
1976	2.7%	3.7%	2.9%	57.9%	27.7%	5.1%
1977	3.2%	3.8%	3.4%	48.7%	33.6%	7.2%
1978	4.4%	4.5%	3.6%	39.9%	36.2%	11.3%
1979	4.8%	4.4%	3.2%	38.1%	37.1%	12.4%
1980	4.5%	4.3%	3.0%	37.5%	35.2%	15.5%
1981	5.8%	4.5%	3.3%	38.6%	32.9%	14.9%
1982	5.6%	4.5%	2.8%	34.5%	33.1%	19.5%
1983	6.9%	4.8%	3.2%	35.4%	28.5%	21.3%
1984	7.2%	4.9%	3.1%	30.9%	28.4%	25.5%
1985	6.0%	4.8%	2.4%	34.1%	30.9%	21.8%
1986	5.8%	4.5%	2.3%	37.4%	28.9%	21.2%
1987	4.9%	4.6%	2.3%	36.3%	33.1%	18.8%
1988	4.8%	4.0%	2.6%	37.7%	35.5%	15.3%
1989	4.8%	4.0%	2.5%	37.7%	33.2%	17.7%
1990	3.8%	3.9%	2.3%	43.0%	35.5%	11.5%
1991	3.4%	3.5%	2.3%	47.4%	34.3%	9.0%
1992	4.0%	3.2%	2.3%	42.8%	35.9%	11.9%
1993	4.6%	3.7%	2.3%	40.7%	34.8%	13.9%
1994	4.8%	4.3%	2.7%	41.2%	34.3%	12.7%
1995	4.4%	4.3%	2.5%	43.1%	33.4%	12.5%
1996	4.7%	4.2%	2.3%	49.5%	28.8%	10.5%
1997	4.0%	3.7%	2.2%	49.0%	30.5%	10.6%
1998	3.1%	3.3%	1.8%	57.0%	25.7%	9.1%
1999	2.2%	2.4%	1.5%	71.2%	18.2%	4.5%
2000	2.4%	1.8%	1.0%	79.1%	12.6%	3.0%
2001	2.6%	2.8%	1.6%	67.8%	20.3%	4.8%
2002	3.0%	3.5%	1.8%	60.3%	24.4%	7.0%
2003	2.7%	3.5%	1.9%	59.3%	23.9%	8.6%
2004	3.6%	3.9%	1.8%	53.3%	26.6%	10.8%
2005	3.9%	3.8%	2.0%	45.8%	31.0%	13.6%
2006	3.8%	4.2%	2.3%	42.9%	33.7%	13.0%
2007	3.7%	4.2%	2.3%	44.4%	28.9%	16.5%
2008	2.8%	3.6%	2.2%	49.3%	31.3%	10.8%
2009	2.7%	3.5%	2.0%	52.9%	30.2%	8.8%
2010	2.7%	3.5%	2.0%	49.0%	27.8%	15.0%
2011	3.2%	3.3%	2.1%	43.3%	28.6%	19.6%
ADDED:

Here is the chart of the data:

Image
Last edited by camontgo on Thu Oct 11, 2012 1:37 pm, edited 1 time in total.
"Essentially, all models are wrong, but some are useful." - George E. P Box

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Thu Oct 11, 2012 1:28 pm

Figure 2 of this article looks at the Market portfolio on a 5X5 grid broken down by size and price (low = growth, high = value): http://www.dfaus.com/2009/05/the-dimens ... -2007.html

If I hadn't told you this was "the market", would you really be referring to this as anything other than a larger, more growth oriented stock portfolio?
The most disciplined investor in the world.

User avatar
SimpleGift
Posts: 3000
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by SimpleGift » Thu Oct 11, 2012 1:48 pm

camontgo wrote:Here is the chart of the data:
Bravo, camontgo! Excellent chart and analysis. Thank you, this directly answers my original question.

I still need to think a bit more about why growth stocks are always over-represented in the large-cap indexes. It certainly has do with the construction methodology of the cap-weighted indexes and the exact definition of a growth stock. The boundary between large-cap growth and value companies appears to be much more permeable over time than I've ever realized.
Cordially, Todd

rkhusky
Posts: 5200
Joined: Thu Aug 18, 2011 8:09 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by rkhusky » Thu Oct 11, 2012 3:59 pm

Jerry_lee wrote:
rkhusky wrote:
Jerry_lee wrote: This fact also throws cold water on the notion that a TSM portfolio is well diversified across all stocks. IT HOLDS all stocks, but is heavily concentrated in the largest, most growth oriented businesses. Thats fine as long as you know it, but its important that you do. The "total" in Total Stock Market is a bit misleading and a marketing contract of Wall Street IMO.
But isn't the reason that those large companies dominate the TSM is because thousands of traders, poring over thousands of pages of business information, set the market price of those companies based on all available information. The traders must feel that those large companies offer the best risk-adjusted future return of their investment dollars. So, by investing in a manner contrary to the market weight, one is making a bet that those thousands of traders are wrong. That seems to me to be true even if the traders are not investing for long term profitability, but only for short term returns. It is the only measure we have for the value of a stock.
No, that isn't why. Big stocks are big and prices are high because the companies are very big and very successful. They don't have the best risk-adjusted returns, just the lowest risk (and lowest expected returns). And investing not according to market weight just means you don't want the risk/return profile of the market portfolio. For all its diversification (# of stocks), a few dozen still make up a large part of the whole. If that isn't what you want your portfolio to look like, then hold other parts of the market in a larger proportion. So long as their expected returns are a function of a risk/return relationship, then you are just altering your overall portfolio profile.

If the world market cap portfolio is 60% equity and 40% fixed income, am I "betting" by holding more or less bonds even if that is appropriate for my situation? The sin isn't in deviating from the market, it is doing so in a concentrated way (holding just a handful of securities) that increases unsystematic/ "bad risks" -- the type we aren't rewarded for taking.

The market portfolio is just another arbitrary allocation that can be attained with really low costs and minimal tax consequences. Don't get me wrong, I love it as a larger/growth holding. I just see many reasons not to be 100% LG.
I think it is a question of how one defines risk and how much weight one gives to risk. The traders obvious goal is to make money and they have, on average, decided that the largest companies are those that provide the best risk-adjusted returns. If there were other companies that provided higher risk-adjusted returns, why wouldn't the money flow to those companies? It appears that those traders who set stock prices have lower tolerance for risk than those that over-weight (with respect to the collective market) more risky stocks. Perhaps it is a difference between looking at past risk and return and trying to predict future risk and return.

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Thu Oct 11, 2012 4:28 pm

rkhusky wrote: I think it is a question of how one defines risk and how much weight one gives to risk. The traders obvious goal is to make money and they have, on average, decided that the largest companies are those that provide the best risk-adjusted returns. If there were other companies that provided higher risk-adjusted returns, why wouldn't the money flow to those companies? It appears that those traders who set stock prices have lower tolerance for risk than those that over-weight (with respect to the collective market) more risky stocks. Perhaps it is a difference between looking at past risk and return and trying to predict future risk and return.
I just think we aren't going to agree on this. Your usage of "traders" isn't really helpful here, because it isn't really what is happeing. Without getting into why there are no securities with the "best risk-adjusted returns" as determined by "traders", the important thing to recognize is simply that people and institutions buy and sell stocks based on available information and collectively, market prices tend to be efficient--meaning that company shares tend reflect their underlying risks most of the time. As investors, we should act accordingly all the time.

Now, that doesn't mean that there is only one source of risk, or we shouldn't seek to understand what those risks are, when those risks tend to show up, and how those risks tend to covary. This is where the FF research helps us. It isn't perfect, but it is far better than almost anything else. For many years, we believed simply that all stocks are the same, and the only risk/return differences exist between equities and fixed income (just one "risk"). Today, we have a richer understanding of the risk entailed with stock and bond investing -- size/value in the former and credit/term in the later. The one thing we can say conclusively is that "one-factor" view of the world is long gone.

So, by all means, buy a "Total Market Index". If this were 1974, your only decision would have to be how much of it relative to t-bills (we hadn't gotten to bond risks yet) for an apporpriate risk/return. Today, we know that other things matter, and other risks generate returns. A TSM fund can still work, but often as a compliment to large and small value stocks, or even conceivably with a large growth index for maximum security diversification and a tilt towards bigger/safer companies.

Just don't put any money in TSM because you believe "that traders have determined this is where the best risk adjusted returns are". That ain't the case.
The most disciplined investor in the world.

User avatar
grayfox
Posts: 4861
Joined: Sat Sep 15, 2007 4:30 am

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by grayfox » Thu Oct 11, 2012 8:45 pm

camontgo wrote: These percentages change over time. Here is the year by year data. Max in LG bucket was 79.1% in 2000, min in LG bucket was 30.9% in 1984. This data could be used to construct a chart similar to that in the OP.

Code: Select all

       SG     SB     SV     LG      LB      LV
1926	1.4%	2.1%	1.7%	37.5%	51.4%	5.9%
1927	1.2%	1.9%	1.6%	49.8%	37.7%	7.8%
1928	0.9%	1.8%	1.8%	45.4%	44.5%	5.5%
1929	0.9%	1.6%	1.4%	49.8%	39.3%	7.0%
1930	0.7%	1.4%	1.1%	64.6%	28.5%	3.8%
1931	0.4%	1.3%	1.0%	70.6%	24.9%	1.8%
1932	0.3%	1.1%	0.8%	76.2%	18.8%	2.9%
1933	0.4%	1.6%	1.2%	68.9%	21.6%	6.3%
1934	0.5%	1.4%	1.1%	52.5%	38.5%	6.0%
1935	0.4%	1.2%	0.9%	50.0%	41.8%	5.7%
1936	0.7%	1.6%	1.1%	51.4%	35.8%	9.3%
1937	1.1%	1.6%	1.2%	43.1%	40.9%	12.1%
1938	0.8%	1.6%	1.2%	62.2%	27.5%	6.7%
1939	0.8%	1.4%	1.1%	51.3%	40.4%	5.0%
1940	0.7%	1.7%	1.2%	60.5%	30.5%	5.5%
1941	0.8%	1.7%	1.4%	59.5%	28.5%	8.1%
1942	0.7%	1.8%	1.4%	63.4%	27.7%	5.1%
1943	1.0%	2.2%	1.7%	62.3%	23.6%	9.2%
1944	1.3%	2.5%	1.8%	55.2%	29.9%	9.4%
1945	1.7%	2.8%	2.0%	42.8%	35.8%	14.9%
1946	2.0%	3.3%	2.6%	37.7%	35.9%	18.5%
1947	1.5%	3.0%	2.2%	40.7%	39.3%	13.2%
1948	1.7%	2.9%	2.1%	43.1%	35.6%	14.6%
1949	1.2%	2.5%	2.1%	58.5%	26.9%	8.7%
1950	1.2%	2.5%	2.2%	50.4%	34.3%	9.4%
1951	1.0%	2.6%	2.3%	50.9%	33.5%	9.7%
1952	1.1%	2.5%	1.9%	55.7%	29.8%	8.9%
1953	0.9%	2.6%	2.0%	56.5%	29.9%	8.1%
1954	0.8%	2.3%	1.9%	60.2%	28.3%	6.6%
1955	0.8%	2.3%	1.9%	60.1%	28.5%	6.3%
1956	0.8%	2.1%	1.7%	59.4%	30.3%	5.6%
1957	0.7%	2.0%	1.7%	64.2%	26.3%	5.0%
1958	0.6%	2.2%	1.7%	62.5%	28.1%	4.8%
1959	0.9%	2.2%	1.8%	60.0%	29.0%	6.2%
1960	0.9%	2.0%	1.7%	54.8%	34.0%	6.6%
1961	1.0%	2.2%	1.8%	62.3%	25.8%	7.0%
1962	1.0%	2.0%	1.8%	49.0%	38.9%	7.3%
1963	1.4%	2.4%	2.0%	60.7%	25.9%	7.6%
1964	1.4%	2.1%	1.9%	58.5%	28.6%	7.5%
1965	1.3%	2.1%	2.1%	55.3%	33.0%	6.1%
1966	1.8%	2.7%	2.2%	50.5%	34.3%	8.6%
1967	1.9%	2.9%	2.6%	51.7%	34.0%	6.8%
1968	3.7%	2.9%	2.8%	43.7%	29.8%	17.0%
1969	3.9%	3.0%	2.4%	38.1%	31.6%	21.0%
1970	2.9%	2.7%	2.1%	40.3%	38.3%	13.7%
1971	2.5%	3.3%	2.7%	50.6%	33.2%	7.8%
1972	2.5%	3.2%	2.5%	54.0%	24.2%	13.7%
1973	2.8%	2.7%	2.2%	54.7%	29.4%	8.2%
1974	2.5%	3.2%	2.2%	62.7%	24.4%	5.0%
1975	2.8%	3.7%	2.5%	65.9%	21.9%	3.3%
1976	2.7%	3.7%	2.9%	57.9%	27.7%	5.1%
1977	3.2%	3.8%	3.4%	48.7%	33.6%	7.2%
1978	4.4%	4.5%	3.6%	39.9%	36.2%	11.3%
1979	4.8%	4.4%	3.2%	38.1%	37.1%	12.4%
1980	4.5%	4.3%	3.0%	37.5%	35.2%	15.5%
1981	5.8%	4.5%	3.3%	38.6%	32.9%	14.9%
1982	5.6%	4.5%	2.8%	34.5%	33.1%	19.5%
1983	6.9%	4.8%	3.2%	35.4%	28.5%	21.3%
1984	7.2%	4.9%	3.1%	30.9%	28.4%	25.5%
1985	6.0%	4.8%	2.4%	34.1%	30.9%	21.8%
1986	5.8%	4.5%	2.3%	37.4%	28.9%	21.2%
1987	4.9%	4.6%	2.3%	36.3%	33.1%	18.8%
1988	4.8%	4.0%	2.6%	37.7%	35.5%	15.3%
1989	4.8%	4.0%	2.5%	37.7%	33.2%	17.7%
1990	3.8%	3.9%	2.3%	43.0%	35.5%	11.5%
1991	3.4%	3.5%	2.3%	47.4%	34.3%	9.0%
1992	4.0%	3.2%	2.3%	42.8%	35.9%	11.9%
1993	4.6%	3.7%	2.3%	40.7%	34.8%	13.9%
1994	4.8%	4.3%	2.7%	41.2%	34.3%	12.7%
1995	4.4%	4.3%	2.5%	43.1%	33.4%	12.5%
1996	4.7%	4.2%	2.3%	49.5%	28.8%	10.5%
1997	4.0%	3.7%	2.2%	49.0%	30.5%	10.6%
1998	3.1%	3.3%	1.8%	57.0%	25.7%	9.1%
1999	2.2%	2.4%	1.5%	71.2%	18.2%	4.5%
2000	2.4%	1.8%	1.0%	79.1%	12.6%	3.0%
2001	2.6%	2.8%	1.6%	67.8%	20.3%	4.8%
2002	3.0%	3.5%	1.8%	60.3%	24.4%	7.0%
2003	2.7%	3.5%	1.9%	59.3%	23.9%	8.6%
2004	3.6%	3.9%	1.8%	53.3%	26.6%	10.8%
2005	3.9%	3.8%	2.0%	45.8%	31.0%	13.6%
2006	3.8%	4.2%	2.3%	42.9%	33.7%	13.0%
2007	3.7%	4.2%	2.3%	44.4%	28.9%	16.5%
2008	2.8%	3.6%	2.2%	49.3%	31.3%	10.8%
2009	2.7%	3.5%	2.0%	52.9%	30.2%	8.8%
2010	2.7%	3.5%	2.0%	49.0%	27.8%	15.0%
2011	3.2%	3.3%	2.1%	43.3%	28.6%	19.6%

Image
Wow, that is an amazing chart!

Also, look at those big peaks --> troughs in LG:
1932-37 76.2% --> 43.1%
1975-84 65.9% --> 30.9%
2000-06 79.1% --> 42.9%

Also, I can't believe how small SV is. So many advise taking huge bets on such a tiny piece of the total market.
In 2000 it was only 1%, now it's doubled its share to 2.1%

yobria
Posts: 5978
Joined: Mon Feb 19, 2007 11:58 pm
Location: SF CA USA

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by yobria » Fri Oct 12, 2012 10:17 am

rkhusky wrote:I think it is a question of how one defines risk and how much weight one gives to risk. The traders obvious goal is to make money and they have, on average, decided that the largest companies are those that provide the best risk-adjusted returns. If there were other companies that provided higher risk-adjusted returns, why wouldn't the money flow to those companies? It appears that those traders who set stock prices have lower tolerance for risk than those that over-weight (with respect to the collective market) more risky stocks. Perhaps it is a difference between looking at past risk and return and trying to predict future risk and return.
Yes, that sums it up nicely. The reason BigCo has a greater market cap than SmallCo is that market participants saw more value there, and drove up the price, making BigCo big. BigCo managed to build 100 oil wells, and SmallCo only built one. If you overweight SmallCo, you're choosing to accept more risk than the market, which has spread its bets equally (all else equal) over all 101 wells. TSM must be the highest return portfolio, adjusting for risk, because it's the most diversified, as my example shows.

Of course this is exactly what the financial industry doesn't want you to know. Nobody gets paid if we all invest in TSM. So they're forced to comb through the historical data, find some pattern that outperformed, and sell a portfolio that overweights that. Problem is, yesterday's free lunch is often today's overvalued asset class. Tech funds in 1999, anyone? To be fair, many industry leaders, even those who profit from tilting (Fama/French/Bernstein/Ferri etc) acknowledge the efficiency of TSM.

User avatar
MickeyBoy
Posts: 36
Joined: Tue Jan 15, 2008 1:39 pm
Location: The Show-Me state

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by MickeyBoy » Fri Oct 12, 2012 10:55 am

A growth-tilted portfolio correlates with the risk factor. That's all.

Hafius500's remark above reminded me of the mention in Ben Stein's Little book of alternative investments that a fund of low-beta stocks - if one existed - would be highly desirable. Would not such a fund fall into the class of Large Value? And would it not represent that asset class better than a fund based on an index structured around the accounting semi-fiction of BtM?

User avatar
baw703916
Posts: 6673
Joined: Sun Apr 01, 2007 1:10 pm
Location: Seattle

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by baw703916 » Fri Oct 12, 2012 11:12 am

MickeyBoy wrote: a fund of low-beta stocks - if one existed - would be highly desirable. Would not such a fund fall into the class of Large Value?
Utilities?
Most of my posts assume no behavioral errors.

DoWahDaddy
Posts: 290
Joined: Sat Jul 02, 2011 9:32 am

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by DoWahDaddy » Fri Oct 12, 2012 11:14 am

Simple stuff, eh? By default the Index 500 is certainly growth heavy. You can always go 75% index 500 and 25% Windsor II to balance (and rebalance). Nice to have that 5bp er core fund. This may or may not be worth the effort. Pick one route and go with it.
Me: 75/25 stocks/bonds | Son: 45/45/10 matchbox/hotwheels/thomas & friends

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Fri Oct 12, 2012 11:38 am

yobria wrote: Yes, that sums it up nicely. The reason BigCo has a greater market cap than SmallCo is that market participants saw more value there, and drove up the price, making BigCo big. BigCo managed to build 100 oil wells, and SmallCo only built one. If you overweight SmallCo, you're choosing to accept more risk than the market, which has spread its bets equally (all else equal) over all 101 wells. TSM must be the highest return portfolio, adjusting for risk, because it's the most diversified, as my example shows.

Of course this is exactly what the financial industry doesn't want you to know. Nobody gets paid if we all invest in TSM. So they're forced to comb through the historical data, find some pattern that outperformed, and sell a portfolio that overweights that. Problem is, yesterday's free lunch is often today's overvalued asset class. Tech funds in 1999, anyone? To be fair, many industry leaders, even those who profit from tilting (Fama/French/Bernstein/Ferri etc) acknowledge the efficiency of TSM.
Is this the reason why Total Market investors are choosing their allocations (because they don't understand how stocks are priced)? I am beginning to wonder if that is the case.

Everything said above is wrong. Let me clean it up:

a) The reason BigCo has a greater market cap (price x # of shares) than SmallCo is because a company's stock price, relative to its earnings or other fundamental metrics, simply reflects the relative degree of risk it carries and the price they have to pay as it relates to the equity component of their cost of capital. That's it. It's not because "traders" have uncovered the best risk-adjusted returns, or because investors "saw more value" there. This couldn't be further from the truth. We know this intuitively because if there were some set of stocks that had the "best value", instead of the fact that all prices simply reflect underlying risk, you would have clear and obvious arbitrage opportunities where sophisticated investors would buy up these stocks, short all the other stocks, and hold the hedge until prices return to underlying risk/return. And evidence from the active community (lets call them "sophisticated investors") are not producing statistically significant alpha adjusted for risk (market, size, and value exposure).

b) If you overweight SmallCo relative to BigCo (or, more likely, overweighting 2000 SmallCo's relative to 500 BigCos in comparison with a LG heavy TSM portfolio), you are simply taking different risks. We know that the small/large dimension is different than the performance of the broad equity market (see 29-42, 65-81, or 00-12 returns for TSM vs SC). There is nothing wrong with this as long as you know what you are doing. Many investors can and should expose themselves to "small risk" instead of concentrating their portfolio in the largest growth stocks. Others will find a LG heavy allocation is preferable (TSM, S&P 500, or Growth Index -- all are acceptably diversified by security and concentrated in terms of risk/return exposure to just BETA).

c) But what TSM doesn't do, no matter how you spin it, is "spread its bets" equally. I don't need to explain this, this particular thread covers this wonderfully. 70% in LC, 20% in MC, and 10% in SC is not "spread equally", no matter how much $ those 500 large companies earn that you have 70% in. Enron was a really big company with very high earnings, several times that of dozens of small companies, but when this company went bust, you still saw the risk of concentrating in one company (no matter how big it was or how much $ it earned). And, of course, because TSM weights a few dozen of these really big companies very heavily, you are significantly exposed to these risks (collectively "beta") no matter how much money they made. One company is still one company, no matter who/how big they are.

TSM is diversified by # of securities, but concentrated in types of securities. In a world where the only risk that bears return is stock/bond, then TSM is as diversified as any equity portfolio. But in OUR WORLD, where returns are a function of stock/bond, small/large, and value/growth, TSM is concentrated in only one dimension of returns.

d) And saying TSM is the highest returning portfolio for its risk means nothing...all diversified portfolios with the same exposure to market, size, and value have the same risk and expected returns. TSMers want you to believe that a Total Stock Index is somehow superior in returns to all other allocations, but it isn't. It simply provides the return for that level of 3F risks: 1.0 market, 0.0 size, and 0.0 value. Other, more tilted portfolios have higher expected returns at the expense of greater exposure to size/value risks. But, thanks to the unique nature of the risks, a tilted portfolio's total risk isn't a linear function of its weighted average risk exposure -- you get a diversification return because of less than perfect correlation between market/size/value. This is no different than the fact that a rebalanced TSM/TBM portfolio's return is greater and risk is less than the weighted average of its components because stock/bond and TERM/CREDIT are not perfectly correlated.

e) I assume Nick is referring to low cost fee-only RIAs when he mentions the "financial industry" doesn't recommend TSM because we wouldn't get paid. But this shows a complete misunderstanding of the services provided by RIAs. Advisors get paid to develop appropriate portfolios for client circumstances, help them navigate all the other issues that come along with wealth (taxes, insurance, estate) by coordinating outside professionals, and making sure they stay with their plan through thick and thin. In this regard, TSM is a tool in the allocation kit. As mentioned before, it is a LG heavy portfolio of 3000 companies with low fees and excellent tax-efficiency. Are there some investors out there for whom a 100% TSM portfolio is appropriate? Probably, but with an almost infinite # of possible portfolios across the size and value spectrum, to assume that one allocation fits all sizes is terribly wrong. In many cases, TSM (and/or S&P 500) is a component of an equally well-diversified (# of securities) portfolio that is spread more intentionally across multiple risk/return dimensions. Just like there isn't a one-size-fits all stock/bond mix, the same applies to the other 4 risk and return factors.
The most disciplined investor in the world.

User avatar
grayfox
Posts: 4861
Joined: Sat Sep 15, 2007 4:30 am

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by grayfox » Fri Oct 12, 2012 12:31 pm

I think that, the whole debate comes down to how diversification is defined. I ran the FF 3-factor regressions for Vanguard Total Stock Market (VTSMX) and Vanguard Small-Cap Value ETF (VBR), using the method described at calculatinginvestor.com in Screencast: Fama-French Regression Tutorial Using R

Code: Select all

               VTSMX        VBR
R^2          0.997        0.97
Intercept    0.0001025   -0.0002115
rmrf         0.9872000    0.981302
smb         -0.0209653    0.6180304
hml          0.0091167    0.4266189
Both had high R^2. The 3-factor model explained 99.7% of VTSMX returns and 97% of VBR returns. In other words the model seems to be a good fit for both the VTSMX and the VBR returns.
Both had ALPHA near zero.

VTSMX had rmrf close to 1, so it captured almost all of market risk premium. But smb and hml were near zero, so VTSMX captured none of small or value risk premium.
VBR also had RMRF close to 1, capturing most of the market risk premium. But smb=0.6180 and hlm=0.4266, so it also captured 60% of small risk premium and 40% of value risk premium.

Under the assumption of the 3-factor model, i.e. that there are 3 risk factors are responsible for equity returns, VBR clearly captured more of the risk factors than VTSMX. So VBR is more diversified, in the sense of spreading your investments over more risk factors. Investors who believe being more diversified is better, would invest in VBR, which is only SV, over VTSMX, which is LV+LB+LG+SV+SB+SG.

But then look at this chart, which shows that SV, by market cap, is only 2% of Total Stock Market.

Image

It may be hard to believe that investors become more diversified by throwing out 98% of the market cap, and concentrating the portfolios in only 2% of the market cap. But under the assumption of the 3 Factor model, that's what it is, if, by diversifiction, you mean exposed to more risk factors.

Edit: By number of stocks, the amount of concentration is not as dramatic. SV=944/3313 = 28.49%. But still you are concentrating in only about 1/4 of the companies out of total U.S. stock market.

Image

Here's another chart that calculatinginvestor posted showing how the number of stocks in each bin varied over time. It looks like LV is the tiniest slice, and most of the stocks are in the the small bins.

Image

So which portfolio is more diversified? all depends on what your yardstick you use.
Last edited by grayfox on Fri Oct 12, 2012 1:32 pm, edited 1 time in total.

hafius500
Posts: 217
Joined: Sat Sep 10, 2011 4:31 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by hafius500 » Fri Oct 12, 2012 12:50 pm

Simplegift wrote:
camontgo wrote:Here is the chart of the data:
Bravo, camontgo! Excellent chart and analysis. Thank you, this directly answers my original question.

I still need to think a bit more about why growth stocks are always over-represented in the large-cap indexes. It certainly has do with the construction methodology of the cap-weighted indexes and the exact definition of a growth stock. The boundary between large-cap growth and value companies appears to be much more permeable over time than I've ever realized.
Camontgo,

How significant are the numbers in the boxes? Can we expect the same patterns in the future? You observed ONE market. Do we see similar patterns and numbers in other markets?

Simplegift,
I still need to think a bit more about why growth stocks are always over-represented in the large-cap indexes. It certainly has do with the construction methodology of the cap-weighted indexes and the exact definition of a growth stock.
"over-represent growth stocks" in relation to ???? What is your benchmark? If sector X is the largest sector, does it mean the real economy "over-represents" X and an appropiate numbers of factories in that sector need to closed?
Or do you say: Growth stocks are always over-represented. Therefore, every market portfolio is always overvalued. Note we talk about rational and efficient markets!

I think all of your questions have been answered:

baw703916 wrote:
If you define growth as the 30% of the total market cap with the lowest BtM and value as the highest 30% by BtM, then a total market fund weighted by market cap will always hold 30% value and 30% growth, simply by mathematical identity
This definition is similar to the academic definitions of growth and value styles I've seen.

Here's a cap-weighted market that "over-represents" large-cap value stocks: 33% Large Value/ 33% Mid Blend/ 33% Small growth.

value (growth) stocks = top (bottom) 33% of the total market cap with the lowest (highest) Price/Book ratios.
Large (small) stocks = ..........................................with the lowest (highest) market cap.

Stock--Market capitalisation--Price/Book Ratio--Style--Weight in cap-weighted Portfolio
A------100 USD----------------1-----------------Large Value--33%
B------ 50 USD----------------2-----------------Mid Blend----16,5%
C------ 50 USD----------------3-----------------Mid Blend----16,5%
D------ 25 USD----------------4-----------------Small Growth-08,25%
E------ 25 USD----------------5-----------------Small Growth-08,25%
F------ 25 USD----------------6-----------------Small Growth-08,25%
G------ 25 USD----------------7-----------------Small Growth-08,25%

The OP asked:
- Is the total market-portfolio growth-tilted? This is a question about factor loadings.
Now you ask:
- Does the total market portfolio "overweight" growth stocks? This is a question about style exposure.
It is the factor loading that defines the expected return. This means that, theoretically, a portfolio of growth stocks can have a postive value loading and a higher expected return than a portfolio of value stocks (or TSM). See the my last post and the data it linked to.
prior username: hafis50

User avatar
camontgo
Posts: 324
Joined: Tue Oct 27, 2009 1:46 pm
Location: Chicago, IL
Contact:

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by camontgo » Fri Oct 12, 2012 2:25 pm

hafius500 wrote:Camontgo,

How significant are the numbers in the boxes? Can we expect the same patterns in the future? You observed ONE market. Do we see similar patterns and numbers in other markets?
I thought it was interesting to look at the market caps of the portfolios formed by the Fama-French ranking method and to see how they vary over time, but I don't have any profound interpretations or forecasts. I expect each portfolio's share of the total market cap will continue to fluctuate. :happy

Some international data is available on the Ken French site, and it would be possible to construct similar charts. I think the data only goes back to 1990, but I may take a closer look at it if I have time this weekend.
"Essentially, all models are wrong, but some are useful." - George E. P Box

User avatar
SimpleGift
Posts: 3000
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by SimpleGift » Fri Oct 12, 2012 2:42 pm

hafius500 wrote:"over-represent growth stocks" in relation to ???? What is your benchmark? If sector X is the largest sector, does it mean the real economy "over-represents" X and an appropriate numbers of factories in that sector need to closed?
Or do you say: Growth stocks are always over-represented. Therefore, every market portfolio is always overvalued. Note we talk about rational and efficient markets!
Apologies, hafius500, this is just imprecise language on my part. Most large-cap stock index funds are classified as "large blend" funds (see Morningstar's categories), leading many investors to assume they are getting a neutral balance of large growth and value companies. However, the value-factor history presented above indicates they are consistently getting an index dominated by large growth companies — thus growth stocks are "over-represented" only in relation to investors' expectations about the index.
Cordially, Todd

User avatar
ClosetIndexer
Posts: 278
Joined: Mon Mar 19, 2012 11:00 pm
Contact:

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by ClosetIndexer » Fri Oct 12, 2012 9:57 pm

Jerry_lee wrote:When I use the Ken French Large Growth/Value indexes on his website along with the S&P 500, I find that the large cap index (S&P 500) sits much closer to growth indexes than value indexes. Here are the specific HmL coefficients since 1926:

FF US Large Growth Index = -0.23
S&p 500 = +0.05
FF US Large Value Index = +0.82

So, you see, there is about 0.28 difference between LG and the S&P 500, and 0.77 difference between S&P 500 and LV. Meaning that the S&P 500 is really sort of a watered down growth index (due to its cap weighting and the fact that companies get to be the biggest because of rising share prices).
Jerry, I agree with much of what you say, and certainly appreciate your expertise and analysis. This particular argument has never made sense to me though. To me, what those numbers show is not that the S&P500 is a watered down growth index, but simply that the FF Large value portfolio has a stronger tilt than the large growth portfolio. The same is true of small value vs. small growth. There are two reasons for this that I know of. First, the value portfolios are more volatile than the growth, so are responsible for more of the movements of the HML factor. Second, the small and large value portfolios are more highly correlated with each other than the growth, and so each captures more of the other half of the overall HML factor (which is an average of small and large HML 'sub-factors'.) Anyway, just because the S&P500's HML weighting is closer to the weighting of that particular growth portfolio than that particular value portfolio doesn't make it a growth index. Its HML weighting is essentially zero - even slightly positive - so it is indeed a large blend index. The fact that the FF LV portfolio is more heavily tilted than the LG portfolio is a separate issue.

Otherwise, it's like you're using the factors to build up to a conclusion, then throwing them out the window to actually draw it. What's ultimately going to matter to risk and return is the factor exposure. IMO it just muddies the waters when we start calling portfolios with positive HML exposure "growth-like". It seems what you're really saying is that the TSM does not capture enough of the HML factor, with its associated premium and volatility, to match your preference, not that a portfolio with +0.2 HML loading is more representative of a 'true' growth/value blend.

Jerry_lee
Posts: 915
Joined: Tue Mar 06, 2012 1:55 pm

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by Jerry_lee » Sat Oct 13, 2012 12:27 pm

ClosetIndexer wrote: Jerry, I agree with much of what you say, and certainly appreciate your expertise and analysis. This particular argument has never made sense to me though. To me, what those numbers show is not that the S&P500 is a watered down growth index, but simply that the FF Large value portfolio has a stronger tilt than the large growth portfolio. The same is true of small value vs. small growth. There are two reasons for this that I know of. First, the value portfolios are more volatile than the growth, so are responsible for more of the movements of the HML factor. Second, the small and large value portfolios are more highly correlated with each other than the growth, and so each captures more of the other half of the overall HML factor (which is an average of small and large HML 'sub-factors'.) Anyway, just because the S&P500's HML weighting is closer to the weighting of that particular growth portfolio than that particular value portfolio doesn't make it a growth index. Its HML weighting is essentially zero - even slightly positive - so it is indeed a large blend index. The fact that the FF LV portfolio is more heavily tilted than the LG portfolio is a separate issue.

Otherwise, it's like you're using the factors to build up to a conclusion, then throwing them out the window to actually draw it. What's ultimately going to matter to risk and return is the factor exposure. IMO it just muddies the waters when we start calling portfolios with positive HML exposure "growth-like". It seems what you're really saying is that the TSM does not capture enough of the HML factor, with its associated premium and volatility, to match your preference, not that a portfolio with +0.2 HML loading is more representative of a 'true' growth/value blend.
CI,

Thanks for the comments. This can get confusing.

Don't be thrown off by the 0.0 of the market. 0.0 isn't "neutral", it's just the market, and positive tilts (coefficients) are in relation to the market portfolio. But we cannot say that the market is "neutral" with respect to value/growth unless it sits equally between LG and LV (or with respect to small/large unless is sits equally between small and large).

We are asking if large cap indexes are more oriented to growth stocks (as measured by a LG index and its relative HmL coefficient) or value stocks (as measured by a LV index and its relative HmL coefficient). What matters here is what the HmL of LV is, what the HmL of LG is, and how close to either the S&P 500 is. If, instead, we want to measure if the S&P 500 is more growth oriented than the market (which is another question), then we would say no -- they are both about 0.0 on HmL. But hopefully we are not forgetting that that is a biased question, as TSM is already more oriented to the LG index than the LV index, as we can see by the fact that 0.0 sits closer to the LG HmL than the LV HmL.

Incidentally, a true "large neutral" portfolio (from the FF 3x3 grid) that excludes both the upper 1/3 of value and bottom 1/3 of growth does sit at about +0.2 HmL, or equal distance between LV and LG and no demonstrated orientation towards one or the another (but I am not aware of an investable version of this index). But true large growth indexes lose this characteristic because they include LG (and LV) stocks and LG companies are higher priced, making up a larger % of the portfolio and pushing them towards the growth end of the spectrum.

PS -- to see this concept visually, see Exibit 3 of this article on page 4: http://www.ifa.com/Media/Images/ThreeFactorsFamaJr.pdf and note the relative position of the S&P and CRSP 1-10 ("market") relative to FF Large Growth and Large Value (DFA)

PSS -- small cap indexes do not tend to have the same growth orientation as large cap indexes do. This can be viewed by looking at SG, SB, and SV indexes in the article above in Exibit 3, or by intuition as we know small indexes sell some higher priced stocks as they become bigger, reducing the growth bias in indexes of all small companies. Some appreciating small stocks become larger (and are sold from the index), others just become more growth oriented (and are held in the index). In a large cap index, either price movement (becoming even larger or more growth oriented) doesn't cause them to leave the index, no amount of largeness or growthness causes a stock to be cast out.
The most disciplined investor in the world.

User avatar
ClosetIndexer
Posts: 278
Joined: Mon Mar 19, 2012 11:00 pm
Contact:

Re: Are Large-Cap Stock Indexes Always Growth-Tilted?

Post by ClosetIndexer » Sat Oct 13, 2012 5:30 pm

Jerry,

I agree that this is essentially a question of what we (somewhat arbitrarily) choose for our mid point. In my view, the market is neutral, so the 30% of stocks that are defined as value are more tilted than the 30% of stocks we define as growth. IE how 'valuey' or 'growthy' something is is simply given by its factor loadings. In yours, the top and bottom thirds (or 30%s) of the market, by BTM, represent 'value' and 'growth', so a point halfway between them (in terms of returns or factor loadings) is neutral.

Whose neutral is "more neutral" depends on the metric you use. Your neutral does have returns that are closer to the midpoint between the growth and value portfolios. However, it has significantly higher correlation to the value portfolio than to the growth. My neutral has approximately the same correlations to the value and growth portfolios, but has returns closer to those of the growth portfolio. Also, while I certainly agree with you that the whole "representative investor" argument to hold TSM doesn't make sense for many investors, it does seem logical to me as reasoning behind calling that 'neutral'.

So again, my way of looking at things is that the market is neutral, and if you sort it and split off an equal number of 'value' and 'growth' stocks, the value ones will be more tilted than the growth. That's primarily because value portfolios are more volatile than growth, so when you take the difference between the two, more of it is captured by the value side. To get returns between those of the value and growth portfolios, you do therefore need a bit of a value factor tilt (around 0.2 to HML), but that will come with corresponding increased volatility and value risk compared to TSM. Fundamentally there's nothing different about calling this the midpoint and saying TSM has less volatility, lower expected return, and less value risk. It just doesn't help me personally to think of it that way.

(As far as saying the same is true for small, I was just pointing out that if you look at the HML loadings of the large and small portfolios, you see the same effect: LV:0.77, LG:-0.27, SV:0.73, SG:-0.24, since 1926/07.)

Post Reply