Small Cap Value: Diversification or Di-Worse-Fication?

[BetaTracker]

On a recent YouTube video, Rob Berger talked to Dr. Karsten Jeske, a longtime market researcher who retired in 2018 and now writes the earlyretirementnow.com blog. One of his articles argues that the statistical significance of overweighting small cap value stocks is less compelling as holding a broadly diversified portfolio of index funds. In fact, according to his research, there's an argument to be made that SCV investing doesn't enhance an overall portfolio's diversification — hence the title he gives it: "Small Cap Value Stocks: Diversification or Di-WORSE-Fication?" He points out that SCV investing isn't new, adding:
"The SCV media blitz relies on the latest narrative that, sure, broad index funds (VTI, ITOT, etc.) are a great and simple way to reach your financial goals. But SCV is an even better way—a “more optimal” way to reach your goals. Allegedly, SCV is the secret sauce for accelerated financial success for astute and enlightened investors. But alas, much of that narrative is hype and false advertising. In today’s post, I want to reiterate the case for simple broad index investing. Let’s take a look…"
In particular, he takes a simple chart showing the S&P 500 vs. a Paul Merriman equal weighted "4 Fund" portfolio (25% each to SCV, SCB, S&P 500, LCV) from 1926 and substitutes the buy-and-hold investing period by starting date rather than end date. It shows a much different picture:
https://earlyretirementnow.com/2024/12/ ... efication/

[White Coat Investor]

Easy to bag on SV after the last 15 years. Nobody was doing it in 2010 (after a decade of 0% returns for the S&P 500) though I assure you. Recency bias anyone?

I think the case for SV tilting can be made quite well from a behavioral standpoint as well as a risk standpoint. And if you liked it in 2010, you should really like it now with the valuation difference between LG and SV at or close to all time highs.

This morning I bought more shares of VTI and more shares of DFSV (along with TLHing a bunch of AVUV to DFSV). So even if Karsten is right and SV tilting is foolish, I've still got more VTI than I'll ever spend.

[Chocolatebar]

I'll just leave this here.



Hopefully you get the point.

[gavinsiu]

I think the SCV might help in certain circumstances. For example, in the last 15 years, small cap value has returned 8.16% real while large cap returned 10.81% real. Small cap value underperform Large cap.

Jump back to 25 years, the small cap value returned 6.9% real while large cap returned 4.9% so the outperformance was enough to persist for 25 years, though obviously this is no indication of the future.

My thoughts are that having a small cap value might have a good diversifying effect. For example during 2000-2009, large cap returned -3.47% real while small cap value returned 5.04%. It may work during high inflation period where stocks and bonds do poorly.

Question, did small cap value do well during 1966-1982?

[BetaTracker]

Research by Fama, French and others argue that small and value stocks hold different risk exposures to the market as compared to large and growth stocks. Since a Total Stock Market Index Fund holds everything, however, isn't factor investing simply an exercise in overweighting those risk factors — and less about exposing a portfolio to a bigger slice of diversification benefits? That seems to be the point of Dr. Jeske's article.

[rkhusky]

Adding a tilt to SV is not diversification, because, as you point out, TSM already includes all the SV stocks at market weight. Adding bonds or international stocks or gold or physical real estate or collectibles to TSM are examples of diversification.

Adding more SV or SG or LV or LG or REIT’s is a tilt, not diversification. In fact, a significant tilt would be concentration, the opposite of diversification.

[dcabler]

Watching this thread in great anticipation that some new argument for or against will finally appear.

[Chocolatebar]

Adding more SV or SG or LV or LG or REIT’s is a tilt, not diversification. In fact, a significant tilt would be concentration, the opposite of diversification.

One of the main arguments I've seen for why an SCV tilt enhances portfolio diversification is that it helps reduce overall portfolio risk by diversifying the specific sources of that risk.

[Kenkat]

It’s de-wors-ification until it has a period of outperformance. Then it becomes a must have, justified by any number of reasonable sounding things. Or you just hold it forever and suffer as I do.

[dcabler]

Adding more SV or SG or LV or LG or REIT’s is a tilt, not diversification. In fact, a significant tilt would be concentration, the opposite of diversification.

One of the main arguments I've seen for why an SCV tilt enhances portfolio diversification is that it helps reduce overall portfolio risk by diversifying the specific sources of that risk.

Yep, as described by "the diversification ratio", for example. If adding a second asset, as long as its correlation is less than 1.0 with the first asset, diversification is improved. Of course, it won't be all that much if the correlation is 0.95. The holy grail is to find a handful of 0 correlation assets, all with high returns. Good luck with that. This definition of diversification becomes asymptotic quickly with a relatively small handful of assets. And by this definition, diversifying sources of risk has nothing to do with improving performance .

Cheers

[White Coat Investor]

Adding a tilt to SV is not diversification, because, as you point out, TSM already includes all the SV stocks at market weight. Adding bonds or international stocks or gold or physical real estate or collectibles to TSM are examples of diversification.

Adding more SV or SG or LV or LG or REIT’s is a tilt, not diversification. In fact, a significant tilt would be concentration, the opposite of diversification.

Depends on if diversification is defined as owning more securities or having a portfolio exposed to more factors.

I mean, do you really think you "own" small value stocks when TSM is only like 3% small value? You have twice that much just in NVIDIA. Talk about concentration.

[Kevin K]

Tyler at Portfolio Charts has a great post about how SCV works in the context of the most efficient (in terms of risk-adjusted return) portfolios a few years ago. Fair warning that in addition to SCV the other useful-in-small-amounts/lethal-in-large assets discussed are those eternal Bogleheads favorites gold and long-term Treasuries.

https://portfoliocharts.com/2021/12/16/ ... ortfolios/

The data also shows that international equities have not been a particularly useful diversifier - but small-cap value has. As Tyler has written elsewhere:

"I like small cap because it provides good diversification away from large cap, not just in correlations but also in company exposure.  Because of cap weighting, the price of large cap funds (or even total market funds) is really driven by a relative handful of mega caps. I also am attracted to the specific combination of VBR with VTI.  They all also include a good amount of mid caps as well, and the resulting combined portfolio has a relatively even spread between small, mid, and large.

It's not so much that I have conviction that small is superior to large.  If anything, I lack conviction that large is superior to small and don't like how a total market fund is less diversified than it seems."

An allocation of half VTI, half VBR seen through the lens of the Morningstar style boxes is almost perfectly split between large, medium and small caps. A different take on "nobody knows nothin'" if you will.

It's no accident that the Golden Butterfly, which has the best risk-adjusted returns of any portfolio on the Portfolio Charts site, uses exactly that allocation.

As for international vs. SCV as a diversifier, the numbers speak for themselves. Quoting Tyler again: "There's more to stock performance than the country it's domiciled in."

https://www.portfoliovisualizer.com/bac ... m9kDry8jkS

[gavinsiu]

I have noticed that SCV have done much better than International as a diversifier, but I wonder about using SCV to cover certain conditions.

For example, how well did international do during the 60's and 70's when inflation is rampant and stock and bond failed to keep up? I don't have any data points.

How well did Japanese small cap value do during the lost decades vs holding international?

[nisiprius]

The data doesn't point strongly in either direction.

1) During the 31-year period of time when small-cap value has been available to investors as a well-structured product intended to capture the factor premium, it has had One Great Shining Moment in 2000-2003.

2) It would be churlish to ignore that, because SCV investors were living the dream: SCV went up when the market went down.

3) It would be particularly churlish to ignore it because the people writing about SCV before 2000 nailed it. They were on the record, in books with model portfolios complete with ticker symbols. Investors got the benefit whether they followed the recommendations of DFA-trained advisors using DFA funds, or the simplified Bill Schultheis "Coffeehouse Portfolio" using Vanguard funds. It was robust and it was called in advance.

4) It has only happened once.

5) Since 2004, the DFA US Small-cap Value Portfolio fund has underperformed Total Stock. But it did so well before that that, that over the whole period the One Great Shining Moment far outweighs the dreary record ever since.





6) The Fama-French data before 1993 is academic data based on cost-free investing. Small-cap value is costly to invest in. The data is based on a rather peculiar method of index construction which presumably makes sense as a study tool for understanding stock behavior. Nevertheless, there are no ETFs or mutual funds that track the Fama-French small-cap-value research portfolio: it's more or less un-investable.

And pure opinion:

Merriman's chart is tendentious, with extra lines on it that are intended to make you see it in a certain way.

[countmein]

Research by Fama, French and others argue that small and value stocks hold different risk exposures to the market as compared to large and growth stocks. Since a Total Stock Market Index Fund holds everything, however, isn't factor investing simply an exercise in overweighting those risk factors — and less about exposing a portfolio to a bigger slice of diversification benefits? That seems to be the point of Dr. Jeske's article.

There are probably no less than 1 million words written on this topic in the forum archives. I politely suggest a search.

[rkhusky]

Depends on if diversification is defined as owning more securities or having a portfolio exposed to more factors.

Factor investing is the opposite of diversification, it’s concentration, you prune the stocks so they all have the same factor characteristics.

[Chocolatebar]

Merriman's chart is tendentious, with extra lines on it that are intended to make you see it in a certain way.

Which 'extra' lines are you referring to?

[Chocolatebar]

Factor investing is the opposite of diversification, it’s concentration.

It seems like you overlooked the points we made and simply repeated what you've already stated. It is possible for concentration to actually enhance portfolio diversification, depending on how you look at it.

Edit: I see that you've updated your comment, but I'm not quite sure I understand what you're saying.

[Kenkat]

2) It would be churlish to ignore that, because SCV investors were living the dream: SCV went up when the market went down.

3) It would be particularly churlish to ignore it because the people writing about SCV before 2000 nailed it. They were on the record, in books with model portfolios complete with ticker symbols. Investors got the benefit whether they followed the recommendations of DFA-trained advisors using DFA funds, or the simplified Bill Schultheis "Coffeehouse Portfolio" using Vanguard funds. It was robust and it was called in advance.

I think an argument could be made that it was not only churlish but borderline insubordinate.

[SantaClaraSurfer]

Imagine three investors with $20,000 per year to invest in taxable (after maxing out their Tax-deferred retirement accounts using a traditional 3 Fund approach).

Investor A, "KeepItSimple": puts 100% in US Total Market

Investor B, "Factorhead": puts $15,000 per year into US Total Market and $5,000 per year into Small Cap Value and does not rebalance.

Investor C "SliceNDicer": maintains a steadily rebalanced 76% allocation into US Total Market and a 24% allocation into a proportional bundle of 3 diversification factor or sector funds at 8% each.

We can all agree that US Total Market will dominate all three portfolios.

Investor A, KeepitSimple, will have a straightforward outcome. After twenty years they will have socked away $400,000 cost basis in a diversified, low cost, total market investment. The biggest risk is the timing of their retirement, and the fate of the US equity market.

Investor B, Factorhead, will likely measure their strategy by whether or not Small Cap Value outperforms US Total Market. If it does not, you can imagine some regret. It if does outperform, then perhaps it was all worth it.

For Investor C, SliceNDicer, it won't be so clear. Every month's new investment will change the complexion of their portfolio slightly based on the performance of the portfolio components. When US Total Market outperforms, Investor C will buy more of the factor/sector investments, and vice versa. Correspondingly, inside of the 24% allocation, when one of the three diversification investments underperforms, they will purchase more of it to bring it to 8% overall.

For Investor C, it will be harder to see the portfolio as a race between the components, especially as, over time, any time US Total Market underperformed in a given year, the three other portfolio components would be used to return it to 76% of the overall allocation. Diversification and rebalancing between multiple investments means, effectively, that all the portfolio components will have contributed to the overall portfolio outcome.

Now, over twenty years, the available invested cost basis for each of the three diversification investments for Investor C in this example would be $32,000 (and $100,000 for Investor B); correspondingly, the invested cost basis of the US Total Market would be about $300,000. Given that, you could fairly argue that both Investor B and C have added needless complication to a portfolio that was destined to rise or fall according the fate of US Total Market, regardless. You could also argue that $32,000 (or even $100,000 in the case of Investor B) in cost basis in an investment after 20 years is a meaninglessly small diversification position and won't make an impact on either investor's retirement outcome.

On the other hand, it takes about 30 minutes to create a spreadsheet that can track and maintain Investor C's strategy. If your goal is diversification as it is best understood, some version of Investor C's strategy is most likely to provide the intended result: comparable portfolio performance with lower risk and volatility. However, it does mean taking the time and effort to build and maintain the strategy; it also means that it won't be 100% clear that one or the other portfolio components significantly outperformed the others as you will have been rebalancing between them incessantly for decades.

It may well be that investing and rebalancing inside of a diversified portfolio pays off in a truly noticeable way only years into retirement; in the meantime, success with any of these strategies is likely best measured by what helps you sleep at night.

The choice probably comes down to the general preferences and outlook of the investor. For myself, I prefer option C, with no expectation of outperformance.

[steadyosmosis]

When I learned that I don't need SCV (nor other tilts) to have a satisfactory portfolio, I abandoned it for simplification.
My VTI shares alone are providing more than I need to pay all of my expenses.

[Taylor Larimore]

Bogleheads:

Whenever I am tempted to "improve" my three total market index funds, I remember what Nobel Laureates recommend:

Douglas Diamond: "Asked what he’ll do with his share of the prize money, Diamond said he’ll probably put it in a total market index fund."

Eugene Fama: "Whether you decide to tilt toward value depends on whether you are willing to bear the associated risk...The market portfolio is always efficient...For most people, the market portfolio is the most sensible decision."

Daniel Kahneman: "Investors shouldn't delude themselves about beating the market. They're just not going to do it. It's just not going to happen."

Harry Markowitz: "A foolish attempt to beat the market and get rich quickly will make one's broker rich and oneself much less so."

Merton Miller: "Most people might just as well buy a share of the whole market, which pools all the information, than delude themselves into thinking they know something the market doesn't"

Paul Samuelson: "The most efficient way to diversify a stock portfolio is with a low-fee index fund. Statistically, a broadly based stock index fund will outperform most actively managed equity portfolios."

William Sharpe: "You may think your opinion is superior, but it pays to be humble, investing in the market rather than trying to beat it."

Robert Shiller: "A portfolio approximating the market may be the most important portfolio."

In my opinion, adding additional small-cap value stocks to a total market index fund is "Di-Worse-Fication".

Best wishes.
Taylor

Jack Bogle's Words of Wisdom: "Don't look for the needle in the haystack, just buy the haystack."

__________________________________________________________________________________________________________

[BetaTracker]

There are probably no less than 1 million words written on this topic in the forum archives. I politely suggest a search.

My apologies, countmein. I did search for posts using the words "Di-Worse-Fication" but didn't find anything. I should've also searched using the author's name, which came up with a discussion on Reddit about Jeske's research. Again, I didn't mean to add a redundant post. Here is the earlier discussion on Bogleheads:
viewtopic.php?t=444223

[rkhusky]

Factor investing is the opposite of diversification, it’s concentration.

It seems like you overlooked the points we made and simply repeated what you've already stated. It is possible for concentration to actually enhance portfolio diversification, depending on how you look at it.

Edit: I see that you've updated your comment, but I'm not quite sure I understand what you're saying.

Start with the market, then remove all the large and midcap stocks, then remove the growth and blend stocks, then remove other stocks based on profitability and investibility and whatever other screens you want, until your stocks are “pure”. How is that diversified? It’s not, it’s concentrated. You may do better with that concentrated bet, no one can tell in advance.

[markus75]

In my opinion, adding additional small-cap value stocks to a total market index fund is "Di-Worse-Fication".

Why you recommended small cap value in 2007 and 2008?
viewtopic.php?p=11146#p11146
viewtopic.php?p=187213#p187213

Has anything other than recent performance changed your view?

[rkhusky]

Adding more SV or SG or LV or LG or REIT’s is a tilt, not diversification. In fact, a significant tilt would be concentration, the opposite of diversification.

One of the main arguments I've seen for why an SCV tilt enhances portfolio diversification is that it helps reduce overall portfolio risk by diversifying the specific sources of that risk.

You are concentrating the risk into a subset of the market, which increases risk, which gives you a chance at higher return.

[Chocolatebar]

Start with the market, then remove all the large and midcap stocks, then remove the growth and blend stocks, then remove other stocks based on profitability and investibility and whatever other screens you want, until your stocks are “pure”. How is that diversified? It’s not, it’s concentrated. You may do better with that concentrated bet, no one can tell in advance.

No one is suggesting you hold only small-cap value stocks. What they’re recommending is tilting your large-cap dominated portfolio towards small-cap value, in order to better balance the overall risk profile. Usually the result looks like S&P500 + SCV domestically (no mid caps or small growth). Like I said before, factor investing helps diversify your portfolio by spreading out independent sources of risk (factors), whereas a MCW portfolio relies solely on market beta.

[BetaTracker]

Whenever I am tempted to "improve" my three total market index funds, I remember what Nobel Laureates recommend:

Much appreciated, Taylor! Years ago, you and pkcrafter helped me to start investing on the old Morningstar forum. Gradually, I became increasingly swayed to try a small-cap value tilt, which then raised questions in my mind how much to overweight small and value. When my employer offered us free access to an advisor, I made the mistake of taking his recommendation to increase our portfolio's tilt. Of course, that was just about the time small value started a lengthy performance slide against the Total Stock Market Index fund.
Since retiring, we've come to realize the importance of simplicity. Our updated financial plan has also reaffirmed that a three-fund index portfolio should more than meet our longer-term retirement needs — and result in a lower standard deviation than our old factor-based portfolio. So, I learned the hard way that factor investing did require us to make a gamble about which parts of the stock market will outperform and, consequently, exposed us to suboptimal investment behavior. (I am also reminded of your wise advice that our asset allocation decision between stocks and bonds was likely to be a bigger driver of long-term portfolio total returns than how much we tweaked our equity allocations.) As you have taught so many of us here, greater portfolio complexity isn't your friend over time.

[rkhusky]

Start with the market, then remove all the large and midcap stocks, then remove the growth and blend stocks, then remove other stocks based on profitability and investibility and whatever other screens you want, until your stocks are “pure”. How is that diversified? It’s not, it’s concentrated. You may do better with that concentrated bet, no one can tell in advance.

No one is suggesting you hold only small-cap value stocks. What they’re recommending is tilting your large-cap dominated portfolio towards small-cap value, in order to better balance the overall risk profile. Usually the result looks like S&P500 + SCV domestically (no mid caps or small growth). Like I said before, factor investing helps diversify your portfolio by spreading out independent sources of risk (factors), whereas a MCW portfolio relies solely on market beta.

The market disagrees that it’s a better balance. Mixing TSM and SCV results in a diversified portion and a concentrated portion.

Theoretically there’s nothing special about SCV. You could tilt TSM to LCV, LCG, SCV or SCG or many other things, to achieve a concentrated portion.

And I’m not sure why someone who believes in factor investing isn’t 100% SCV. After all, the long-only SCV fund gets 100% of the market beta, along with the SV factors. Diluting SCV with TSM shows a lack of faith in the factors.

[GP813]

Small cap value has good fundamentals even if the market isn't recognizing it. AVUV is trading at a low p/e, excellent price to sales, etc. Once people start buying companies based on fundamentals again and not "stories", small cap value should do well.

[Beensabu]

SCV often behaves/performs differently than the broad market, because it is a subsection of the broad market.

SCV is subject to risks beyond market risk, because it is a subsection of the broad market. These additional risks may or may not result in additional return.

Tilting to a subsection of the broad market is considered diversification when it outperforms the broad market.

Tilting to a subsection of the broad market is considered di-worse-ification when it underperforms the broad market.

If you're constantly going to compare a tilt to the broad market, just invest in the broad market and don't tilt.

[tomas123]

The post seems to take aim at people that believe that SCV is some kind of magical sauce that will consistently make your portfolio outperform. Not entirely sure that's a large crowd.

Also.. all the t-stat analysis and charts, are they even valid? Seems odd to calculate the t-stat on the difference between the returns of a 1 fund portfolio (S&P 500) and a 4 fund portfolio (SCV, SCB, LCV, S&P 500) where the s&p 500 is included in both. The independence assumption is definitely violated, affecting both average difference and variability.

[rkhusky]

Small cap value has good fundamentals even if the market isn't recognizing it. AVUV is trading at a low p/e, excellent price to sales, etc. Once people start buying companies based on fundamentals again and not "stories", small cap value should do well.

I thought SCV had low P/E and higher risk because it has bad fundamentals.

[nisiprius]

...
Also.. all the t-stat analysis and charts, are they even valid?
...

I've been railing about this (occasionally) for years, and never have gotten an answer from the people who cite them.

As nearly as I can tell, the people who quote multiple "T-stats" in articles are not using Bonferroni corrections and do not understand what they are or why they matter. At least, they never say whether or not they have used them.

There are legitimate statistical techniques for making multiple simultaneous tests, but I've never had the statistical chops to use them, or know when it is OK to use them.

[Nathan Drake]

Tyler at Portfolio Charts has a great post about how SCV works in the context of the most efficient (in terms of risk-adjusted return) portfolios a few years ago. Fair warning that in addition to SCV the other useful-in-small-amounts/lethal-in-large assets discussed are those eternal Bogleheads favorites gold and long-term Treasuries.

https://portfoliocharts.com/2021/12/16/ ... ortfolios/

The data also shows that international equities have not been a particularly useful diversifier - but small-cap value has. As Tyler has written elsewhere:

"I like small cap because it provides good diversification away from large cap, not just in correlations but also in company exposure.  Because of cap weighting, the price of large cap funds (or even total market funds) is really driven by a relative handful of mega caps. I also am attracted to the specific combination of VBR with VTI.  They all also include a good amount of mid caps as well, and the resulting combined portfolio has a relatively even spread between small, mid, and large.

It's not so much that I have conviction that small is superior to large.  If anything, I lack conviction that large is superior to small and don't like how a total market fund is less diversified than it seems."

An allocation of half VTI, half VBR seen through the lens of the Morningstar style boxes is almost perfectly split between large, medium and small caps. A different take on "nobody knows nothin'" if you will.

It's no accident that the Golden Butterfly, which has the best risk-adjusted returns of any portfolio on the Portfolio Charts site, uses exactly that allocation.

As for international vs. SCV as a diversifier, the numbers speak for themselves. Quoting Tyler again: "There's more to stock performance than the country it's domiciled in."

https://www.portfoliovisualizer.com/bac ... m9kDry8jkS

It is not true that International stocks are not a particularly useful diversifier; quite the opposite in fact.

The latest data from the Cederburg paper shows that those with only single country allocations had 100% higher risk of failure compared to those with a global allocation.

While recent history may show that international hasn't been a good diversifier over the last 30 years, we need to not rush to making conclusions based on a fairly limited sample size. The broadest set of data from the Cederburg paper is far more insightful than the past few decades alone.

While exposure to other risk factors, like Value, can also enhance portfolio long term outcomes, as we have seen in the recent YTD performance - US SCV is performing even worse than US TSM; while International Stocks (both TSM and Value performing about equally) have made significant gains. So in this scenario, the domestic SCV market did not diversify as strongly as International Stocks did.

[Nathan Drake]

The data doesn't point strongly in either direction.

1) During the 31-year period of time when small-cap value has been available to investors as a well-structured product intended to capture the factor premium, it has had One Great Shining Moment in 2000-2003.

2) It would be churlish to ignore that, because SCV investors were living the dream: SCV went up when the market went down.

3) It would be particularly churlish to ignore it because the people writing about SCV before 2000 nailed it. They were on the record, in books with model portfolios complete with ticker symbols. Investors got the benefit whether they followed the recommendations of DFA-trained advisors using DFA funds, or the simplified Bill Schultheis "Coffeehouse Portfolio" using Vanguard funds. It was robust and it was called in advance.

4) It has only happened once.

5) Since 2004, the DFA US Small-cap Value Portfolio fund has underperformed Total Stock. But it did so well before that that, that over the whole period the One Great Shining Moment far outweighs the dreary record ever since.





6) The Fama-French data before 1993 is academic data based on cost-free investing. Small-cap value is costly to invest in. The data is based on a rather peculiar method of index construction which presumably makes sense as a study tool for understanding stock behavior. Nevertheless, there are no ETFs or mutual funds that track the Fama-French small-cap-value research portfolio: it's more or less un-investable.

And pure opinion:

Merriman's chart is tendentious, with extra lines on it that are intended to make you see it in a certain way.

My arguments:

1) Again - why are we only focusing on a singular country?

2) Why not analyze Emerging Markets or International Developed Markets? The Value effect is live and well internationally.

3) Furthermore - why cherry pick a start date for US SCV? People investing in US SCV since 2019 have seen their portfolio have a higher average value over the entire sample, only recently coming under pressure from this year's US market correction. If you are making regular contributions to a structured plan, you are constantly rebalancing into assets at more ideal pricing entry points, so your overall performance may be better.

4) Small Cap Value can be "bursty" but so is the overall TSM premium over Bonds. It is not uncommon to have very brief periods of very high performance, regardless of the Equity Risk Premium or the Value Risk premium.

[enad]

Read the speech that Jack Bogle Bogle gave back in 2002, titled The Telltale Chart in which he discusses the reversion to the mean

[breakfastinbed]

The data doesn't point strongly in either direction.

1) During the 31-year period of time when small-cap value has been available to investors as a well-structured product intended to capture the factor premium, it has had One Great Shining Moment in 2000-2003.

2) It would be churlish to ignore that, because SCV investors were living the dream: SCV went up when the market went down.

3) It would be particularly churlish to ignore it because the people writing about SCV before 2000 nailed it. They were on the record, in books with model portfolios complete with ticker symbols. Investors got the benefit whether they followed the recommendations of DFA-trained advisors using DFA funds, or the simplified Bill Schultheis "Coffeehouse Portfolio" using Vanguard funds. It was robust and it was called in advance.

4) It has only happened once.

5) Since 2004, the DFA US Small-cap Value Portfolio fund has underperformed Total Stock. But it did so well before that that, that over the whole period the One Great Shining Moment far outweighs the dreary record ever since.





6) The Fama-French data before 1993 is academic data based on cost-free investing. Small-cap value is costly to invest in. The data is based on a rather peculiar method of index construction which presumably makes sense as a study tool for understanding stock behavior. Nevertheless, there are no ETFs or mutual funds that track the Fama-French small-cap-value research portfolio: it's more or less un-investable.

And pure opinion:

Merriman's chart is tendentious, with extra lines on it that are intended to make you see it in a certain way.

Just want to point out that your analysis is cherry-picked. You say DFSVX underperformed total market if you exclude 15% of the sample? OK. If you exclude a single year (1998) DFSVX has crushed total market since inception. Or since 2004, DFSVX again easily beats total market...if you exclude 2017-2020. After all, DFSVX not only beat the total market from 1993-2004, it also beat the total market from 2005 through 2016, and from 2021 until now. It's not that SCV has spurts of out-performance but overall lags; historically, the total market outperforming SCV for a prolonged stretch has been the anomaly.

"SCV went up when the market went down...It has only happened once."
Well, other than 2021-2022, when DFSVX beat the total market by 30% over two years.

So, fund A beat fund B if you exclude the 5 year period where fund B beat fund A by an average of 20.7%/year but don't exclude the 4 year period where fund A beat fund B by an average of 13.6%/year?

Of course, DFSVX, and many other assets, absolutely crush the total market if you exclude a single stock, Nvidia, for even a fraction of the sample.


You well know that total market returns are driven by a subset of holdings having runs of out-performance. You thus know you can't go back and remove certain years and keep others; wouldn't that be described as tendentious?

Images for purposes of illustration:

[White Coat Investor]

And I’m not sure why someone who believes in factor investing isn’t 100% SCV. After all, the long-only SCV fund gets 100% of the market beta, along with the SV factors. Diluting SCV with TSM shows a lack of faith in the factors.

Yes, it's important not to tilt more than you believe.

[jjj_22]

And I’m not sure why someone who believes in factor investing isn’t 100% SCV. After all, the long-only SCV fund gets 100% of the market beta, along with the SV factors. Diluting SCV with TSM shows a lack of faith in the factors.

Yes, it's important not to tilt more than you believe.

Indeed. So much discussion about all this from both sides. My personal best estimate is that it's about a coin flip either way. So I'm 50/50

[Jack56]

The data that I have seen seems to show that small cap has not produced alpha, consistent outperformance, in around 40 years. Prof. Damodaran published something on this recently. If it is diversification you are looking for you can find it in total market portfolio. If you think you can do better than average by moving in to/out of small cap at the optimal time, e.g., after it has under/over performed for X years, good luck.

[the_wiki]

.

And I’m not sure why someone who believes in factor investing isn’t 100% SCV. After all, the long-only SCV fund gets 100% of the market beta, along with the SV factors. Diluting SCV with TSM shows a lack of faith in the factors.

If you believe stocks will outperform bonds, why hold bonds?

[breakfastinbed]

.

And I’m not sure why someone who believes in factor investing isn’t 100% SCV. After all, the long-only SCV fund gets 100% of the market beta, along with the SV factors. Diluting SCV with TSM shows a lack of faith in the factors.

If you believe stocks will outperform bonds, why hold bonds?

Yup. It makes zero sense to invest in bonds. Total bond has a long-term negative real return over the past 50 years. No one in their right mind should hold them.

...until you look at your portfolio as a whole. Let's use a more extreme example, 25+ year zero coupon treasuries. Testfolio gives 64 years of data; stocks out-performed zeros 56% of the time, by an average margin of 24.4%, and overall returned a CAGR 3x higher (6.18% vs 2.05%). Of course, 27% of the time zeros beat stocks by double-digits, and had several brief periods where they smoked stocks - over 1981-1986 zeros outperformed stocks by 3.5x, and 2x over 2007-2011. Etc. Even though stocks are a better investment in isolation, the two work better together than in comparison.

We have 35 years of SCV data. We can debate whether total market or SCV is better. But half of those years, one of the two out-performed the other by double digits. SCV has had higher returns, but also is more volatile, by a good amount; combining the two (50/50 for example) gave the same return as SCV (so, higher than total market) but reduced the excess volatility by 2/3rds.

And as they always say: past results =/= future returns...

[TextualChocolate]

50/50 Large Cap Growth / Small Cap Value during accumulation is what the academic research supports as well as the backtesting: https://www.portfoliovisualizer.com/bac ... NDYpwoYAuQ

[Random Walker]

One of the main arguments I've seen for why an SCV tilt enhances portfolio diversification is that it helps reduce overall portfolio risk by diversifying the specific sources of that risk.

Yep, as described by "the diversification ratio", for example. If adding a second asset, as long as its correlation is less than 1.0 with the first asset, diversification is improved. Of course, it won't be all that much if the correlation is 0.95. The holy grail is to find a handful of 0 correlation assets, all with high returns. Good luck with that. This definition of diversification becomes asymptotic quickly with a relatively small handful of assets. And by this definition, diversifying sources of risk has nothing to do with improving performance .

Cheers

The correlation of long only SV funds with TSM is going to be very high, but I believe that is because the SV funds have an exposure to market beta of 1.0 or slightly greater. But when looking at the pure factors in isolation, I believe the correlations are very low. Last I looked they were something on the order of
Size to market 0.4
Value to market 0.1
Size to value 0.1

I believe most all of our portfolios are dominated by market beta. Even the typical BH 60% TSM/40% BND portfolio has 85-90% of its risk wrapped up in market beta. So if the goal is to create a more efficient portfolio, we need to diversify away from market beta. That means first estimating an expected return for the TSM/BND portfolio and second creating a portfolio tilted to SV with similar expected return AND lower overall equity allocation. So perhaps the 60/40 total market portfolio becomes a 40/60 SV/BND portfolio. The 40/60 portfolio now has its risks more evenly spread across market beta, size, value, term: a move in the direction of risk parity. The key to improving overall portfolio efficiency is the decrease in overall equity exposure TOGETHER WITH the tilt.
Looking backward over the last 15 years or so though, this strategy has lost significantly relative to the total markets approach. In absolute terms it’s done ok because there is still a lot of market beta in the portfolio.

Dave

[Random Walker]

One of the main arguments I've seen for why an SCV tilt enhances portfolio diversification is that it helps reduce overall portfolio risk by diversifying the specific sources of that risk.

Yep, as described by "the diversification ratio", for example. If adding a second asset, as long as its correlation is less than 1.0 with the first asset, diversification is improved. Of course, it won't be all that much if the correlation is 0.95. The holy grail is to find a handful of 0 correlation assets, all with high returns. Good luck with that. This definition of diversification becomes asymptotic quickly with a relatively small handful of assets. And by this definition, diversifying sources of risk has nothing to do with improving performance .

Cheers

If one can create a more efficient portfolio that has same expected return as more conventional portfolio, has smaller SD, and cuts both the good right tail and the bad left tail of the potential return distribution, he may well have a portfolio that is more robust during withdrawal phase. This portfolio may well have significantly higher likelihood of success and possibly support a higher withdrawal rate.

Dave

[Random Walker]

Adding a tilt to SV is not diversification, because, as you point out, TSM already includes all the SV stocks at market weight. Adding bonds or international stocks or gold or physical real estate or collectibles to TSM are examples of diversification.

Adding more SV or SG or LV or LG or REIT’s is a tilt, not diversification. In fact, a significant tilt would be concentration, the opposite of diversification.

Depends on if diversification is defined as owning more securities or having a portfolio exposed to more factors.

I mean, do you really think you "own" small value stocks when TSM is only like 3% small value? You have twice that much just in NVIDIA. Talk about concentration.

Moreover, the LG exposure in TSM offsets the SV exposure, so no net exposure to any factor other than market factor.

Dave

[Random Walker]

SCV often behaves/performs differently than the broad market, because it is a subsection of the broad market.

SCV is subject to risks beyond market risk, because it is a subsection of the broad market. These additional risks may or may not result in additional return.

Tilting to a subsection of the broad market is considered diversification when it outperforms the broad market.

Tilting to a subsection of the broad market is considered di-worse-ification when it underperforms the broad market.

If you're constantly going to compare a tilt to the broad market, just invest in the broad market and don't tilt.

Sometimes a negative result actually supports an argument. In this case where people are trying to diversify across unique and independent risk factors, the lesser performance of SV perhaps does show exposure to unique and independent risks. The risks have just shown up over the recent past.

Dave

[Random Walker]

Read the speech that Jack Bogle Bogle gave back in 2002, titled The Telltale Chart in which he discusses the reversion to the mean

I believe a problem with that speech is that he was discussing active funds. But here’s the point I really want to make. What mean are we reverting to?!! If size and value are additional risk factors, then the mean expected return for SV funds should be higher than the mean expected return of TSM. Markets price risk and we all believe risk and expected return are related. So it’s only rational to have different mean expected returns to revert to.

Dave

[enad]

Read the speech that Jack Bogle Bogle gave back in 2002, titled The Telltale Chart in which he discusses the reversion to the mean

I believe a problem with that speech is that he was discussing active funds. But here’s the point I really want to make. What mean are we reverting to?!! If size and value are additional risk factors, then the mean expected return for SV funds should be higher than the mean expected return of TSM. Markets price risk and we all believe risk and expected return are related. So it’s only rational to have different mean expected returns to revert to.

Dave

Whether it's rational or irrational to invest in tilts is up to the individual investor. Sometimes any tilt including SCV works, and sometimes it doesn't. Simplicity is King right?