Research Affiliates article on size premium

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
garlandwhizzer
Posts: 1882
Joined: Fri Aug 06, 2010 3:42 pm

Research Affiliates article on size premium

Post by garlandwhizzer » Tue Dec 23, 2014 2:16 pm

RA argues in new paper that current research suggests the size premium is not now statistically significant and hasn't been so since 1981 when it was first described. This applies to both US and all international markets. Interesting--shortly after it was discovered it disappeared or at least weakened to the point of being statistically insignificant. Comments?

Garland Whizzer

[Content in excess of copyright fair use removed by admin LadyGeek. Here's the article: Busting the Myth About Size]

dkturner
Posts: 1348
Joined: Sun Feb 25, 2007 7:58 pm

Re: Research Affiliates article on size premium

Post by dkturner » Tue Dec 23, 2014 2:44 pm

Interesting. I went back and compared the Russell 1000 and Russell 2000 for the period 1979-2013 and found the same result, namely no meaningful difference in annualized returns between large and small stocks for the last 35 years.

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

Re: Research Affiliates article on size premium

Post by richard » Tue Dec 23, 2014 2:58 pm

What if a large company were split, on paper only, into two small companies? Suppose there is no change in operations, and imagine that one of the small companies booked all the cash flows on even-numbered days of the month, and the other one accounted for all the cash on odd days. In this scenario, it would be most surprising if the small companies both delivered higher returns than the original large company. Yet the size premium is precisely based on the expectation that small-cap stocks will outperform large-cap stocks!
The usual response to this is hand waving about more companies means more diversification and smaller means more vulnerable, i.e., riskier. Anyone have a better counter to this argument?

exeunt
Posts: 907
Joined: Tue Jan 08, 2008 6:54 pm

Re: Research Affiliates article on size premium

Post by exeunt » Tue Dec 23, 2014 3:14 pm

dkturner wrote:Interesting. I went back and compared the Russell 1000 and Russell 2000 for the period 1979-2013 and found the same result, namely no meaningful difference in annualized returns between large and small stocks for the last 35 years.
The R2K is a flawed index that has lost about 1% to 2% a year to its horrendously inefficient reconstitution. You can verify this yourself by running the R2K through a FF regression: there's significant size and value, but also nasty negative alpha.
richard wrote:
What if a large company were split, on paper only, into two small companies? Suppose there is no change in operations, and imagine that one of the small companies booked all the cash flows on even-numbered days of the month, and the other one accounted for all the cash on odd days. In this scenario, it would be most surprising if the small companies both delivered higher returns than the original large company. Yet the size premium is precisely based on the expectation that small-cap stocks will outperform large-cap stocks!
The usual response to this is hand waving about more companies means more diversification and smaller means more vulnerable, i.e., riskier. Anyone have a better counter to this argument?
I don't believe size is a significant factor, but this argument is a bit disingenuous. Small and large companies are different along many different characteristics, including growth prospects, financing, competitive advantages, etc. Obviously, in the example above, if all things are equal, size doesn't necessarily matter.

The funny thing is many academics and practitioners know size is a fairly useless standalone factor, but it persists in large part thanks to DFA's marketing efforts.

User avatar
saltycaper
Posts: 2650
Joined: Thu Apr 24, 2014 8:47 pm
Location: The Tower

Re: Research Affiliates article on size premium

Post by saltycaper » Tue Dec 23, 2014 3:18 pm

I appreciate papers that take extreme anomalies like those mentioned here re: 1930s and compare results with/without. I can't believe more people don't do this. It's like basing recommendations on holding long bonds because they've outperformed some stock styles for the past 30 years.

Anyway, I wonder what the data would show comparing large value vs. small value. Small growth might really be dragging the small-cap world down.
Quod vitae sectabor iter?

pauliec84
Posts: 412
Joined: Wed Feb 09, 2011 11:29 am

Re: Research Affiliates article on size premium

Post by pauliec84 » Tue Dec 23, 2014 3:29 pm

It is standard practice to do this in academia. It is called "winsorizing" (http://en.wikipedia.org/wiki/Winsorising)

An issue though is that you have to remove (or modify as in winsorizing) the both the positive and negative outliers. If not you bias the data in the other direction.

dkturner
Posts: 1348
Joined: Sun Feb 25, 2007 7:58 pm

Re: Research Affiliates article on size premium

Post by dkturner » Tue Dec 23, 2014 3:42 pm

exeunt wrote:
dkturner wrote:Interesting. I went back and compared the Russell 1000 and Russell 2000 for the period 1979-2013 and found the same result, namely no meaningful difference in annualized returns between large and small stocks for the last 35 years.
The R2K is a flawed index that has lost about 1% to 2% a year to its horrendously inefficient reconstitution. You can verify this yourself by running the R2K through a FF regression: there's significant size and value, but also nasty negative alpha.
Understood. If I compare the Russell 3000 and the Russell 1000 I also see virtually identical annualized returns over the period 1979-2013. It was my understanding that the poor performance of the Russell 2000 index was due to front-running of stocks moving into and out of the 2000 index. Since the 3000 index includes all of the stocks in both the L-C 1000 and S-C 2000 the front-running issue should either disappear or be greatly minimized shouldn't it?

exeunt
Posts: 907
Joined: Tue Jan 08, 2008 6:54 pm

Re: Research Affiliates article on size premium

Post by exeunt » Tue Dec 23, 2014 3:52 pm

dkturner wrote:Understood. If I compare the Russell 3000 and the Russell 1000 I also see virtually identical annualized returns over the period 1979-2013. It was my understanding that the poor performance of the Russell 2000 index was due to front-running of stocks moving into and out of the 2000 index. Since the 3000 index includes all of the stocks in both the L-C 1000 and S-C 2000 the front-running issue should either disappear or be greatly minimized shouldn't it?
The effects of any adds/deletes from/to the R1K should be neutralized, but the effects of the other adds/deletes would still exist (i.e. stocks that have become too small for the R2K are dropped completely and microcap stocks that have become big enough for the R2K get added).

User avatar
JoMoney
Posts: 5258
Joined: Tue Jul 23, 2013 5:31 am

Re: Research Affiliates article on size premium

Post by JoMoney » Tue Dec 23, 2014 3:54 pm

When Benjamin Graham described the "Bargain-Issue Pattern in Secondary Companies." in The Intelligent Investor, it was not presented as a strategy for passive investors. Passive investors were warned to explicitly avoid junk bonds and small-caps, that they should limit their selections to larger companies and high-grade bonds. For more enterprising investors though, it was pointed out that there have been pendulum swings in the markets sentiment about these securities and it can be at times over-done on the downside presenting an opportunity for those who have a sense of the intrinsic value. But this was not something suggested to be attempted for a passive investor, or something that should be expected as a persistent "risk premium". But it was pointed out that someone who did invest in a broadly diversified group of these securities in the long-run may have been able to achieve an "average" outcome.
Benjamin Graham wrote:...In other words, an investor who bought all such issues at their offering prices might conceivably fare as well, in the long run, as one who limited himself to first-quality securities; or even somewhat better. But for practical purposes the question is largely irrelevant. Regardless of the outcome, the buyer of second-grade issues at full prices will be worried and discommoded when their price declines precipitately. Furthermore, he cannot buy enough issues to assure an “average” result, nor is he in a position to set aside a portion of his larger income to offset or “amortize” those principal losses which prove to be permanent. Finally, it is mere common sense to abstain from buying securities at around 100 if long experience indicates that they can probably be bought at 70 or less in the next weak market.

...Such situations are not for the inexpert investor; lacking a real sense of values in this area, he may burn his fingers. But there is an underlying tendency for market decline in this field to be overdone; consequently the group as a whole offers an especially rewarding invitation to careful and courageous analysis. ...
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

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

Re: Research Affiliates article on size premium

Post by richard » Tue Dec 23, 2014 3:55 pm

exeunt wrote:<>
richard wrote:
What if a large company were split, on paper only, into two small companies? Suppose there is no change in operations, and imagine that one of the small companies booked all the cash flows on even-numbered days of the month, and the other one accounted for all the cash on odd days. In this scenario, it would be most surprising if the small companies both delivered higher returns than the original large company. Yet the size premium is precisely based on the expectation that small-cap stocks will outperform large-cap stocks!
The usual response to this is hand waving about more companies means more diversification and smaller means more vulnerable, i.e., riskier. Anyone have a better counter to this argument?
I don't believe size is a significant factor, but this argument is a bit disingenuous. Small and large companies are different along many different characteristics, including growth prospects, financing, competitive advantages, etc. Obviously, in the example above, if all things are equal, size doesn't necessarily matter.

The funny thing is many academics and practitioners know size is a fairly useless standalone factor, but it persists in large part thanks to DFA's marketing efforts.
So small is different because it has many different characteristics? This sounds suspiciously like begging the question.

User avatar
JoMoney
Posts: 5258
Joined: Tue Jul 23, 2013 5:31 am

Re: Research Affiliates article on size premium

Post by JoMoney » Tue Dec 23, 2014 4:09 pm

dkturner wrote:Interesting. I went back and compared the Russell 1000 and Russell 2000 for the period 1979-2013 and found the same result, namely no meaningful difference in annualized returns between large and small stocks for the last 35 years.
The same situation can be found with the S&P500 compared to the Wilshire4500 index.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

User avatar
nisiprius
Advisory Board
Posts: 36034
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Research Affiliates article on size premium

Post by nisiprius » Tue Dec 23, 2014 5:38 pm

pauliec84 wrote:It is standard practice to do this in academia. It is called "winsorizing" (http://en.wikipedia.org/wiki/Winsorising)

An issue though is that you have to remove (or modify as in winsorizing) the both the positive and negative outliers. If not you bias the data in the other direction.
The problem is that accurate financial statistics are so fat-tailed, so rich in extreme values... to the point where the extreme values are the data... that it's not clear that the procedure would be valid for financial data. CF for example studies that emphasize the importance of investing in the total market, because a large fraction of the return of the total market is generated by a very few single exceptional companies--so if you hold only 50 stocks, say, you have great uncertainty due to possibility of missing or including those superstocks in your sample.

Cheerleaders for stocks sometimes have a tendency to act as if, say, 1929-1936 should be treated as an outlier that should be discarded. On the one hand, 1929-1936 obviously shouldn't be discarded. On the other hand, there's a certain validity to saying that stock markets' going to zero (2 of the 23 stock markets studied by DImson & al) shouldn't count, not because it didn't happen, but because in that kind of societal disruption you may be worrying about other things, like personal survival.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
in_reality
Posts: 4529
Joined: Fri Jul 12, 2013 6:13 am

Re: Research Affiliates article on size premium

Post by in_reality » Wed Dec 24, 2014 3:22 am

richard wrote:
What if a large company were split, on paper only, into two small companies? Suppose there is no change in operations, and imagine that one of the small companies booked all the cash flows on even-numbered days of the month, and the other one accounted for all the cash on odd days. In this scenario, it would be most surprising if the small companies both delivered higher returns than the original large company. Yet the size premium is precisely based on the expectation that small-cap stocks will outperform large-cap stocks!
The usual response to this is hand waving about more companies means more diversification and smaller means more vulnerable, i.e., riskier. Anyone have a better counter to this argument?
No, I would agree with that.

Consider their conclusion:
We are not arguing that investors should completely abandon small stocks. Small stocks are more volatile than large stocks, and they receive considerably less attention from sell-side analysts. Consequently, small stocks are more likely to be mispriced. The major anomalies are, in fact, stronger in the small-cap sector. Small stocks are more attractive as an alpha pool to be fished by skillful active managers and exploited by rules-based value and momentum strategies.
I think the article does well to consider that stocks do go under, drop out and return a real loss. Other research I've read suggest that value stocks return a premium because they are more difficult to price. For growth stocks, you can more easily just subtract expect earnings from the current price in a downturn. For value stocks, it's difficult to price them effectively because if the company survives it will be worth much more than if it goes under and is liquidated. Not only is there uncertainty for survival but there is additional uncertainty for what the assets of the company can actually fetch in down economic times especially since the assets are likely to have been acquired for a specific purpose and may not be readily repurposed.

What's interersting (or perhaps confusing) about the RAFI methodology is that their small cap funds tend to have a fair amount of mid cap growth. Obviously their screens on fundamentals are returning relatively valuey companies in relatively unvaluey sectors so things aren't a traditional value tilt. Their small cap funds end up having 8-9% more growth than their total market funds do. OK I admit to not knowing why this is and using the Schwab Russell (Fundamental) series anyway. Bad bad me (investing in something I don't completely understand) perhaps but I do understand how the funds tilt value after growth outperformance and tilt less value after value outperformance. Why small/mid are different than large is beyond me a little though.

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Research Affiliates article on size premium

Post by Robert T » Wed Dec 24, 2014 4:19 am

.
My take on the size premium.

1. The CRSP data show the small cap premium to be largest in decile 10 (microcaps)
  • Annualized returns: 89 years to June 30, 2014
    CRSP Deciles: Annualized Returns
    Decile 1: …9.36%
    Decile 2: .10.75%
    Decile 3: .11.04%
    Decile 4: .11.14%
    Decile 5: .11.68%
    Decile 6: . 11.57%
    Decile 7: . 11.73%
    Decile 8: . 11.78%
    Decile 9: . 11.73%
    Decile 10: 13.36%

    Source: Bridgeway Funds Annual Report

    I had earlier looked into the delisting bias debate. http://www.bogleheads.org/forum/viewtop ... 81#p172381 My understanding at the time was that it explains some, but doesn’t explain all of the decile 10 blip.

    In my view – there is a risk story – a cut and paste from a previous post.
    Robert T wrote:What’s the source of the CRSP10 non-linearity? Transaction costs indeed play a role but IMO CRSP10 stocks have fundamentally higher risk. Consider the following table. The first column is the average company size by decile (derived from Ibbotson Yearbook), the second column is a transpose of work by Gutierrez that estimates the average company size by Moody credit rating. When transposed onto the deciles it largely maps out as in the second column below. The third column is the percentage default rates by Moody rating. Default rates in decile 10 appear to be substantially higher than the other deciles reflecting much higher risk. While this refers to bond default, the higher fundamental risk of these companies IMO is also reflected in the cost of equity capital and hence equity returns – leading to the CRSP10 non-linearity relative to the other deciles.
    Image
2. Bridgeway Ultra-Small Company Market has matched the annualized returns of Decile 10 since inception of the fund 7/31/1997 (although with significant annual deviations)
  • Since inception: 7/31/1997 to 9/30/2014 Annualized Return
    11.09% - Bridgeway Ultra-Small Company Market (BRSIX)
    11.58% - CRSP10

    So despite, transaction costs, delisting biases, etc, Bridgeway has managed to get live returns close to CRSP10 returns for the full period, more so than DFA.

    1998-2014 YTD - Annualized Return
    .6.4% = Vanguard 500
    .9.8% = DFA US Microcap
    11.7% = Bridgeway Ultra-Small Co. Market
3. Over the full period since 1982 the annualized return difference between large caps and small caps has been much smaller
  • 1982-2014 YTD – Annualized return
    11.6% = Vanguard 500
    12.3% = DFA US Micro
This is the time period of focus in the article.


4. Small caps can underperform for long periods, and the FF small cap premium has been less consistent than the FF value premium
  • According to the linked Fama-French paper, there’s a 10 percent chance of no small cap premium over a 50 year time period, and 14 percent chance over 25 years; a higher level of uncertainty than the value premium. http://us.dimensional.com/pdf/Volatilit ... emiums.pdf
5. There has been a significant ‘small cap’ premium in value stocks
  • The ‘small cap’ premium appears to be larger in value stocks than the average across all stocks.

    1982-2014 YTD Annualized returns
    12.5% = Large Value
    14.2% = Small Value
    +1.7% = “Small cap premium in value stocks”

    Large Value = Ibbotson Large Value 1982-1993, DFA Large Value 1994-2014
    Small Value = Ibbotson Small Value 1982-1993, DFA Small Value 1994-2014

    The difference between the RAFI Fundamental Small Value and Large Value seems to be even bigger.

    I recall reading this earlier article by Arnott, when he was editor of the Financial Analysts Journal http://www.researchaffiliates.com/Produ ... _Value.pdf . The article tries to disentangle to size and value effect, and he concluded in the article that "...Two-thirds of the return associated with the size effect is evidently attributed to the P/S [Price/Sales] component of market capitalization- a value effect – whereas only one-third is a true size effect." I would just note though, that the value (HmL) load on size (SmB) is small (statistically insignificant) - according to Asness, replicating Fama-French https://www.aqr.com/cliffs-perspective/ ... ng-the-way . So not sure about the exact answer.
Bottom line – my take:

- If you want to tilt to small caps (but don’t want an explicit value-growth tilt), Bridgeway Ultra-Small Company Market (BRSIX) offers a good choice to do this and probably gets closer to CRSP10 returns than any other fund currently available. Would use only in tax-advantaged account.

- Small value stocks seem to have offered a significant ‘small cap premium’. For those wanting an explicit tilt to value, in my view, the Powershares RAFI Fundamental Pure Small Value (PXSV) offers a good choice to do this and if using this fund, probably not necessary to also add BRSIX.

- Given the lower consistency in the size premium (with a greater chance of longer periods of no-premium, than the value premium – according to the earlier linked Fama-French article), I have a value load target double that of my size load target (0.4 vs. 0.2), or put another way, a size load target half my value load target (perhaps a reflection of my relative confidence in the respective premiums). I have had the same targets for last 12 years, not plans to change now.

Obvious no guarantees.

Robert
.

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

Re: Research Affiliates article on size premium

Post by richard » Wed Dec 24, 2014 6:53 am

Robert T wrote:<>In my view – there is a risk story – a cut and paste from a previous post.
Robert T wrote:What’s the source of the CRSP10 non-linearity? Transaction costs indeed play a role but IMO CRSP10 stocks have fundamentally higher risk. Consider the following table. The first column is the average company size by decile (derived from Ibbotson Yearbook), the second column is a transpose of work by Gutierrez that estimates the average company size by Moody credit rating. When transposed onto the deciles it largely maps out as in the second column below. The third column is the percentage default rates by Moody rating. Default rates in decile 10 appear to be substantially higher than the other deciles reflecting much higher risk. While this refers to bond default, the higher fundamental risk of these companies IMO is also reflected in the cost of equity capital and hence equity returns – leading to the CRSP10 non-linearity relative to the other deciles.
If it's a risk story and credit rating is a reasonable measure of risk, why not invest in stocks of companies with low credit ratings? That would seem a more direct method than choosing small stocks.

Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Research Affiliates article on size premium

Post by Rodc » Wed Dec 24, 2014 8:58 am

richard wrote:
What if a large company were split, on paper only, into two small companies? Suppose there is no change in operations, and imagine that one of the small companies booked all the cash flows on even-numbered days of the month, and the other one accounted for all the cash on odd days. In this scenario, it would be most surprising if the small companies both delivered higher returns than the original large company. Yet the size premium is precisely based on the expectation that small-cap stocks will outperform large-cap stocks!
The usual response to this is hand waving about more companies means more diversification and smaller means more vulnerable, i.e., riskier. Anyone have a better counter to this argument?
I'm not sure the real world is all that similar to this hypothetical example. In this example they seem to share management at the highest levels, share financing, etc. just as they did as a big company, so I would think the risk profile would not really match two independent small companies. How much that matters I have no idea.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.

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

Re: Research Affiliates article on size premium

Post by richard » Wed Dec 24, 2014 10:07 am

Rodc wrote:
richard wrote:
What if a large company were split, on paper only, into two small companies? Suppose there is no change in operations, and imagine that one of the small companies booked all the cash flows on even-numbered days of the month, and the other one accounted for all the cash on odd days. In this scenario, it would be most surprising if the small companies both delivered higher returns than the original large company. Yet the size premium is precisely based on the expectation that small-cap stocks will outperform large-cap stocks!
The usual response to this is hand waving about more companies means more diversification and smaller means more vulnerable, i.e., riskier. Anyone have a better counter to this argument?
I'm not sure the real world is all that similar to this hypothetical example. In this example they seem to share management at the highest levels, share financing, etc. just as they did as a big company, so I would think the risk profile would not really match two independent small companies. How much that matters I have no idea.
Assume a company with 10 very similar factories. Split it into two companies with 5 factories each. Make the two new smaller companies as similar as possible. Higher return for the two smaller companies?

Larger doesn't necessarily mean anything about management quality. A larger company may be able to pay more, but the relation between pay and performance is not very reliable. Larger doesn't necessarily mean easier to finance. Things like cash flow coverage, debt to equity, etc. would seem at least as important.

garlandwhizzer
Posts: 1882
Joined: Fri Aug 06, 2010 3:42 pm

Re: Research Affiliates article on size premium

Post by garlandwhizzer » Wed Dec 24, 2014 1:37 pm

Robert T's post is as always informative, interesting, and insightful. He seems to suggest that there is a risk tradeoff for CRSP decile 10 stocks, the only decile that shows really significant outperformance. Risk as measured by bond yields, a good measure of risk, is higher in that decline, explaining their outperformance not as a free lunch but rather as a reward for talking on more risk. The delisting figures of small caps relative to large caps support this increased risk.

I would like to add my own personal view of the small cap value premium's history. I believe that the entire outperformance of SCV over the last 90 years can be traced to a few discrete market events in which there were dramatic bubbles in valuation of large cap stocks, especially LCG. The first ending in 1929 when shoe shine boys were giving Joe Kennedy hot stock tips. The last ending in 2000 with the massive overvaluation of LCG, a Shiller PE10 of 44, never seen before or since. The third (less significant) was the era of the Nifty Fifty where investors believed all they had to do to assure outperformance was load up on the same 50 selected large cap growth stocks.

Large cap growth has historically tended to be the place where bubbles inflate. In the aftermath of the bubble's bursting, LCG tanks severely while SCV, which has tended to suffer years of neglect during the inflation of the bubble, outperforms not for months but for years afterward as relative valuations revert to the mean. This, in my opinion is the source of the SCV premium. It has little to do with SCV and a lot to do with LCG bubbles. After this process of unwinding is over in the aftermath, SCV does not historically outperform at all. In fact in terms of risk adjusted returns it tends to underperform both LCB and MCB.

In my opinion the market today, 14 years after the last LCG bubble and 5 years after the financial crisis, has more than corrected the excesses of LCG. Relative valuations of LC versus SC and of G versus V have in my opinion corrected fully. In fact I believe that fundamentally SC is overvalued today relative to LC due to the popularity of factor chasing funds and ETFs . PE ratios of SC relative to LC remain near historical extremes even after the 2014 partial SC correction. In short, I do not expect SC or SCV to outperform in near term starting from here. When the next LCG bubble inflates and real relative value returns to SCV, I'll get fully on board the SCV train then, but not until.

Garland Whizzer

dkturner
Posts: 1348
Joined: Sun Feb 25, 2007 7:58 pm

Re: Research Affiliates article on size premium

Post by dkturner » Wed Dec 24, 2014 2:04 pm

garlandwhizzer wrote:Robert T's post is as always informative, interesting, and insightful. He seems to suggest that there is a risk tradeoff for CRSP decile 10 stocks, the only decile that shows really significant outperformance. Risk as measured by bond yields, a good measure of risk, is higher in that decline, explaining their outperformance not as a free lunch but rather as a reward for talking on more risk. The delisting figures of small caps relative to large caps support this increased risk.

I would like to add my own personal view of the small cap value premium's history. I believe that the entire outperformance of SCV over the last 90 years can be traced to a few discrete market events in which there were dramatic bubbles in valuation of large cap stocks, especially LCG. The first ending in 1929 when shoe shine boys were giving Joe Kennedy hot stock tips. The last ending in 2000 with the massive overvaluation of LCG, a Shiller PE10 of 44, never seen before or since. The third (less significant) was the era of the Nifty Fifty where investors believed all they had to do to assure outperformance was load up on the same 50 selected large cap growth stocks.

Large cap growth has historically tended to be the place where bubbles inflate. In the aftermath of the bubble's bursting, LCG tanks severely while SCV, which has tended to suffer years of neglect during the inflation of the bubble, outperforms not for months but for years afterward as relative valuations revert to the mean. This, in my opinion is the source of the SCV premium. It has little to do with SCV and a lot to do with LCG bubbles. After this process of unwinding is over in the aftermath, SCV does not historically outperform at all. In fact in terms of risk adjusted returns it tends to underperform both LCB and MCB.

In my opinion the market today, 14 years after the last LCG bubble and 5 years after the financial crisis, has more than corrected the excesses of LCG. Relative valuations of LC versus SC and of G versus V have in my opinion corrected fully. In fact I believe that fundamentally SC is overvalued today relative to LC due to the popularity of factor chasing funds and ETFs . PE ratios of SC relative to LC remain near historical extremes even after the 2014 partial SC correction. In short, I do not expect SC or SCV to outperform in near term starting from here. When the next LCG bubble inflates and real relative value returns to SCV, I'll get fully on board the SCV train then, but not until.

Garland Whizzer
Very good summary. My personal belief is similar to yours in that I also believe that the SCV rally, which began in 2000, was a once in a generation phenomenon, much like the rally in corporate bonds which began in 2009. If one is fortunate enough to identify these rare imbalances the rewards can be very substantial. If you miss them you're unlikely to improve on the returns of the classic three fund portfolio. Right now there don't seem to be any serious imbalances, although foreign stock markets bear watching.

Stray the course! 8-)

User avatar
nedsaid
Posts: 9697
Joined: Fri Nov 23, 2012 12:33 pm

Re: Research Affiliates article on size premium

Post by nedsaid » Wed Dec 24, 2014 2:31 pm

Garland, as normal you have stimulated a lot of thought on this forum. It appears that a lot of the arguments about the size premium are centered on how good the academics' data really is. If it can be shown that their data or methodology were flawed (cherry picking) then the case for the small and value premiums are very flawed.

The reason that I have believed in the performance factors talked about by the academics is that the factors resonate with what I know about human nature and markets. It makes sense to me that there would be a small premium because of higher risk and greater potential for growth. It also makes sense from the standpoint of these companies being less known and less followed by Wall Street. Perhaps this is more a philosophical argument than an argument over data. Perhaps this is about how individuals see the world.

Unfortunately, we are human and tend to see the things we want to see. For example, I always marveled at how much technical analysts could see from the wiggles and squiggels of a line on a graph. Sometimes you wonder if it all is a big of a Rohrschach test, seeing what we want to see from inkblots.

I have read a lot of this stuff, as much as I can without my eyes glazing over. The sense that I get is that Fama and French and others looked at the data and went to wherever the data took them. I don't think they had big preconceived ideas before they started this research. Now that these ideas have been published and discussed so much it would be much harder to look at the data objectively. Mr. Bogle would never be convinced of performance factors no matter how much data was thrown at him. He would say that it is all period dependent. Those who believe in the factors would sift the data until they got what they wanted. The problem of bias is a real one and not easy to get around.

The de-listing effect and survivor's bias in small-cap stocks is a really great point. Somehow I have to believe that the academics factored this in to their analysis. It is too big of a thing to miss. On the other hand, don't indexes exhibit survivor bias themselves? If you really wanted an objective index of market performance, shouldn't you include companies that went to zero? I wonder if this is a bit of a red herring.

But I certainly don't believe there is bad faith on the part of the academics. Indeed, I have believed what they have said and made some adjustments to my portfolio. The academics have quantified and better articulated what the investment community has known for years. Size, value, momentum, quality, illiquidity, and other factors have been known about and discussed for years. This is another reason I don't poo-poo the academic research.

I am not a market expert and I certainly am not a quant. But I have owned investments for years and watched how they performed in real life situations. What I have observed is consistent with what the academics are now saying. But again, my observations are in narrative form. My only data is from Quicken and I can give folks a good idea of how my investments have actually performed. There is nothing super precise about what I am saying. I have digested a lot of experience and knowledge over the years and after a while you get sort of a "feel" about things. But I can say that I have seen all the factors that the academics have talked about operate in the real world of investing.

I have made arguments for dividend investing for example and I have done what others have done and exclude the extremes from my thinking. For example, Nisiprius pointed out that dividends did not provide downside protection during the depression and also during the 2008-2009 financial crisis. My counterpoint was there wasn't much out there that did provide the downside protection and that in times of panic, we see the baby, the bathwater, and even the bathtub flying out the window. I further pointed out that the best investment approaches don't work all the time.

I think behind what Garland and others are saying is that human behavior has a way of throwing a clink into the works. Things seem to happen in the one way you never would have expected. I applaud Garland for his comments of Large Cap Growth bubbles and relative valuations of the different parts of the stock market. He might be on to something here that maybe the small value premium is really the reverse of Large Cap Growth bubbles. What I like about his comments is that he is looking at valuations as I do.

The other thing I take out of all of this is that perhaps investors ought to take another look at the mid-cap area of the market. You get most of the benefits of small caps with less of the problems, but you might sacrifice a bit in return. The small cap indexes and funds carry a lot of mid-caps in them anyway.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 9697
Joined: Fri Nov 23, 2012 12:33 pm

Re: Research Affiliates article on size premium

Post by nedsaid » Wed Dec 24, 2014 3:16 pm

It seems to me that this discussion of factors really boils down to this. Can I as an individual investor capitalize on the behavioral mistakes of other investors and beat the market by doing so?

The small and value premium is really a way of saying to not pile into the most popular and followed areas of the market. Buy into what is unloved and underfollowed.

Momentum is a way of taking advantage of investor behavior. What is popular tends to get more popular. . .up to a point. Where small and value are contrarian in nature, momentum is the ultimate follow the crowd strategy.

Quality is a way of saying that the market will pay up for great companies. As long as the companies you invest in pursuing this strategy meet their numbers, you can get a premium. But these type of companies are very vulnerable to disappointments. Another follow the crowd strategy.

The problem is that while human nature doesn't change, people aren't stupid. Perhaps enough people have figured out the small value premium that now the premium will go away for a while. What I count on is that such people will lose interest and again gravitate towards what is popular.

I suppose indexing is a follow the crowd strategy albeit at very low costs. But a broad index will cover both the popular and unpopular stocks in it. Maybe it is a hybrid philosophy between follow the crowd and a contrarian viewpoint.

Here are other questions to ask. Should I try to follow the crowd or be a contrarian? Or should I try to blend the two. Or should I be contrarian sometimes and follow the crowd at other times according to market conditions? Or should I just throw in the towel and stick with the market weighted indexes?
A fool and his money are good for business.

stlutz
Posts: 4547
Joined: Fri Jan 02, 2009 1:08 am

Re: Research Affiliates article on size premium

Post by stlutz » Wed Dec 24, 2014 5:28 pm

First off, let me be the first to say that I hope everyone realizes a big Christmas-factor premium tomorrow!

A few thoughts:

First, "small" has always has a definitional problem because it's not independent of "value." That is, small = price and value = price/book (or some other fundamental value). So, if you drop the price, it becomes both smaller and more valuey. I think other RA research has shown that if you instead define small based on revenues, book value etc., the small premium also basically goes away. In other words, much of what has been claimed as a "small" premium may just be an indirect way to get value exposure.

Eric Falkenstein actually researched the question of the returns of stocks with junk credit ratings vs. stocks with investment-grade credit ratings and the later group crushed the former. He at least would argue that small stocks do not outperform because they have bad credit ratings.

The disproportionate outperformance of small value vs. large value actually suggests market/informational inefficiency. The problem that most of the research has is that it's simply not based upon what information people were actually trading on at the time. What we know is that a) Decile 10 stocks are primarily the domain of individual investors (even more so historically), and that b) individual investors didn't have a way to easily access real data on such companies prior to the the SEC EDGAR database coming online in the mid-90s. As such, the research on small value very much suggests that most people were trading on the wrong types of information (e.g. broker tips instead of fundamental ratios). As such, it is reasonable to ask whether the market we have today actually resembles the one of the 1960s--especially when it comes to small stocks.

One of the big problems with academic backtesting is that they base it on databases that are designed to be a complete and clean as possible based on everything we know today. That is very different from basing them on what was known at the time investors were making decisions. So, they include data on smallcap companies that were collected years and even decades after the fact. The have the correct financials for Enron and not the falsified numbers (was it a value stock or a growth stock at the time?) They've got all of the data that was corrected later by companies for reasons other than fraud as well as fixes for simple data entry errors.

There are financial databases that show the information that was actually in them at the time (i.e. showing multiple dimensions of time). However, they only date back to the late 80s at the furthest and the academic community does not make use of them. Studies that use these databases still show a value premium, just a smaller one than academics do.

The big value of papers like this for readers of this forum is to illustrate that data is not just data. Financial history is not a controlled experiment. Because a bunch of math gets used and because numbers are presented precise to two decimal points does not make more qualitative questions (e.g. "To what extent was the market of the 1960s like today's) irrelevant.

User avatar
JoMoney
Posts: 5258
Joined: Tue Jul 23, 2013 5:31 am

Re: Research Affiliates article on size premium

Post by JoMoney » Wed Dec 24, 2014 6:05 pm

nedsaid wrote:... It appears that a lot of the arguments about the size premium are centered on how good the academics' data really is. If it can be shown that their data or methodology were flawed (cherry picking) then the case for the small and value premiums are very flawed. ...
Even if their data are good, there are issues with the premises they start with. The EMH school and the belief that that the market is efficiently pricing "risk premiums" has lead to all sorts of strategies that have ended with some the biggest economic calamities. There's no saying that the companies represented in small-cap stocks today are anything like the small-caps of 50 years ago. Given all the various changes in the economy, finance, and technology I'd wager they're quite different. Even if there is a real risk factor associated with smaller-cap stocks, there's no saying that the markets sentiment and peoples opinions on how to value that "risk" hasn't changed, especially when the popular view becomes that it's a route to out-performance. As Buffett wrote at the end of the junk bond debacle :
Warren Buffett in 1990 Berkshire Annual Letter wrote: [...]selling investment bankers pointed to the "scholarly" research of academics, which reported that over the years the higher interest rates received from low-grade bonds had more than compensated for their higher rate of default. Thus, said the friendly salesmen, a diversified portfolio of junk bonds would produce greater net returns than would a portfolio of high-grade bonds. (Beware of past-performance "proofs" in finance: If history books were the key to riches, the Forbes 400 would consist of librarians.)

There was a flaw in the salesmen's logic - one that a first- year student in statistics is taught to recognize. An assumption was being made that the universe of newly-minted junk bonds was identical to the universe of low-grade fallen angels and that, therefore, the default experience of the latter group was meaningful in predicting the default experience of the new issues. (That was an error similar to checking the historical death rate from Kool-Aid before drinking the version served at Jonestown.) ...
... and again during the home loan crisis ...
Warren Buffett in 2008 Berkshire Annual Letter wrote:...Indeed, the stupefying losses in mortgage-related securities came in large part because of flawed, history-based models used by salesmen, rating agencies and investors. These parties looked at loss experience over periods when home prices rose only moderately and speculation in houses was negligible. They then made this experience a yardstick for evaluating future losses. They blissfully ignored the fact that house prices had recently skyrocketed, loan practices had deteriorated and many buyers had opted for houses they couldn’t afford. In short, universe “past” and universe “current” had very different characteristics. But lenders, government and media largely failed to recognize this all-important fact.
Investors should be skeptical of history-based models. Constructed by a nerdy-sounding priesthood using esoteric terms such as beta, gamma, sigma and the like, these models tend to look impressive. Too often, though, investors forget to examine the assumptions behind the symbols. Our advice: Beware of geeks bearing formulas.
The idea that because something performed well in the past, that it will do so in the future is problematic. Especially when it's based on assumptions that someone else is pricing the risk for you, and believing you can blindly buy at whatever price because the market is efficiently pricing "risk premiums" for you. This goes for the larger stock market as well, there have been problems that creep in when the common belief is that stocks are always better forever and ever..
Warren Buffett on Risk Premiums wrote:... Stocks usually yield a little more, but that isn't ordained. Every once in a while, stocks will get very cheap, but it isn't ordained in scripture that this is so. Risk premiums are mostly nonsense. The world isn't calculating risk premiums.

The best book prior to Graham was written by Edgar Lawrence Smith in 1924 called Common Stocks as Long Term Investments. It was a study that evaluated how bonds compared to stocks in various decades of the past. There weren't a whole lot of publicly traded companies back then. He thought he knew what he was going to find. He thought that he'd find that bonds outperformed stocks during periods of deflation, and stocks outperformed during inflationary times. But what he found was that stocks outperformed the bonds in nearly all cases. John M. Keynes then enumerated the reasons that this was so. He said that over time you have more capital working for you, and thus dividends would grow higher. This was novel information back then and investors then went crazy and started buying stocks for these higher returns. But then they started to get crazy, and no longer really applied the sound tactics that made the reasons given in the book true. Be careful that when you buy something for a sound reason, make sure that the reason stays sound.
... Quote from Ben Graham: “You can get in more trouble with a sound premise than an unsound premise because you'll just throw out the unsound premise”.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

User avatar
nedsaid
Posts: 9697
Joined: Fri Nov 23, 2012 12:33 pm

Re: Research Affiliates article on size premium

Post by nedsaid » Wed Dec 24, 2014 10:15 pm

JoMoney, you raised excellent points. This is why you are one of my favorite posters on the forum.
The world sometimes really does change and old models become obsolete.

This is why I tilt but I don't overdo it. My heart sinks when I see people who have all their equities in small cap value. Even when coupled with a large allocation of Treasuries (the Larry portfolio), I just think this is reckless. I don't get these extreme portfolios. Yes, the data backs up the Larry portfolio but something in me says that one needs more diversification than that.

Investing is a very humbling experience. Once one thinks he has it all figured out, a rare event occurs in the market, and he marvels at how little he really knows. Market surprises really do happen.

I also have to allow for the fact that I might be wrong. Thus I do a whole lot of things in my portfolio. I tilt small value but I own large growth stocks. I am value oriented but I own aggressive investments. So I diversify across asset classes, diversify across factors, and frankly diversify against my own stupidity.
A fool and his money are good for business.

Park
Posts: 513
Joined: Sat Nov 06, 2010 4:56 pm

Re: Research Affiliates article on size premium

Post by Park » Thu Dec 25, 2014 12:28 am

I may very well being cynical, but the small cap premium, especially the small cap value premium, is much easier to obtain for individual investors than institutional investors. Am I surprised that Research Affiliates concluded that the small premium is not of importance?

User avatar
in_reality
Posts: 4529
Joined: Fri Jul 12, 2013 6:13 am

Re: Research Affiliates article on size premium

Post by in_reality » Thu Dec 25, 2014 12:36 am

Park wrote:I may very well being cynical, but the small cap premium, especially the small cap value premium, is much easier to obtain for individual investors than institutional investors.
I don't think they said small value is unimportant, rather that small overall isn't.
We are not arguing that investors should completely abandon small stocks. Small stocks are more volatile than large stocks, and they receive considerably less attention from sell-side analysts. Consequently, small stocks are more likely to be mispriced. The major anomalies are, in fact, stronger in the small-cap sector. Small stocks are more attractive as an alpha pool to be fished by skillful active managers and exploited by rules-based value and momentum strategies.
Am I surprised that Research Affiliates concluded that the small premium is not of importance?
You tell us. Are you?

lazyday
Posts: 3297
Joined: Wed Mar 14, 2007 10:27 pm

Re: Research Affiliates article on size premium

Post by lazyday » Thu Dec 25, 2014 5:03 am

Park wrote:I may very well being cynical,
I'd think that if RA's work suffered because of financial interests, others would call them out.

But I don't know how many highly regarded academics in the field aren't involved with DFA, RA, AQR, etc and have similar incentives.

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Research Affiliates article on size premium

Post by Robert T » Thu Dec 25, 2014 5:36 am

garlandwhizzer wrote:In my opinion the market today, 14 years after the last LCG bubble and 5 years after the financial crisis, has more than corrected the excesses of LCG. Relative valuations of LC versus SC and of G versus V have in my opinion corrected fully. In fact I believe that fundamentally SC is overvalued today relative to LC due to the popularity of factor chasing funds and ETFs . PE ratios of SC relative to LC remain near historical extremes even after the 2014 partial SC correction. In short, I do not expect SC or SCV to outperform in near term starting from here. When the next LCG bubble inflates and real relative value returns to SCV, I'll get fully on board the SCV train then, but not until.
It may be true that the expected returns, over the near term (7-10 years) of small caps are lower than large caps given current valuations, as per GMO and RAFI forecasts. It may be possible to market time asset classes based on current valuations, but history seems to suggest that accurately timing asset class rotations is a fairly tough proposition. While I currently use the GMO 7-year forecasts for rebalancing within 5x25 bands around fixed target allocation, I am not sure this will add much/anything. In my view a better proposition than attempting asset class rotation/timing is to set a fixed allocation, which will do okay in different market conditions - and stick with it. For example - consider the period - 1985-1999, where small value underperformed (re: the data Taylor periodically posts and reminds us of).

1985-1999 – Annualized Return

15.8% = Dimensional US Small Value
18.7% = Dimensional US Large Value
18.9% = S&P 500

Over this 15 year time period small value lagged the S&P500 by 3 percent per year (annualized), or put another way, the S&P 500 returns were 20 percent higher. Lets assume that going forward this may be the relative performance of the S&P500 vs. Small Value (based on current forecasts from GMO and RAFI). One response is sell SV and buy S&P500. If indeed there was this level of relative under-peformance for 15 years, how many would actually stick with SV, while watching (and reading) how everyone else is doing better by simply buying the S&P500. It seems to only take a couple of years of underperformance (let alone 15 years) for human tendency to look/shift elsewhere - often to the detriment of long-term returns. As we see in 2000-2013 the ordering of asset class returns was reversed (as per below) but over the full period (1985-2013) SV outperformed, and seemed to provide more consistent returns over the two periods. In my view, this latter effect is the benefit of factor diversification (market, small, and value exposure), it prevents the larger extremes from concentration in a single factor (e.g. beta only in this example). But with media hype generally focused on beta (the market/S&P500), when the S&P500 outperforms its hard to stay the course with a more diversified portfolio.

2000-2013 – Annualized Return

14.8% = Dimensional US Small Value
..7.9% = Dimensional US Large Value
..3.6% = S&P 500

IMO a better proposition than asset class (or factor) rotation (i.e. selling SV and buying the S&P500), is to structure a portfolio (diversify across factors) in a way that can reduce tracking error when beta (the market/S&P500) does well, to increase the likelihood of staying to course. In my view adding the "momentum factor' can help to do this - now that there is an available option (MTUM that tracks the MSCI USA momentum series). Consider the commonly referenced 30% S&P500, 30% large value, 40% small value portfolio. This portfolio would have outperformed the S&P500 in 2000-2013 (9.6% vs. 3.6% annualized return), but underperformed in 1985-1999 (17.8% vs. 18.9% annualized return). If we replace the S&P500 with MSCI Momentum, the portfolio keeps up in 1985-1999 (19.4% vs. 18.9% annualized return).

2000-2013 – Annualized Return
..3.6% = S&P500
..9.6% = 30% S&P500, 30% Dimensional Large Value, 40% Dimensional Small Value
10.2% = 30% MSCI Momentum, 30% Dimensional Large Value, 30% Dimensional Small Value

1985-1999 – Annualized Return
18.9% = S&P500
17.8% = 30% S&P500, 30% Dimensional Large Value, 40% Dimensional Small Value
19.4% = 30% MSCI Momentum, 30% Dimensional Large Value, 30% Dimensional Small Value

Just to note: The above result is not exclusive to the Dimensional series. Using 30% MSCI Momentum, 30% Russell MidCap Value, and 40% RAFI Fundamental Small Value gives similar results. 11.6% annualized for 2000-2013, and 18.6% annualized for 1985-1999. For the full periods the simulated annualized returns were 15.1% with a standard deviation of 18.6, compared to 14.9% return with standard deviation of 18.9 with the Dimensional series.

Even if momentum underperforms going forward, as it has done in Japan, then the value funds will likely perform better as they have negative momentum loads, and as they have done in Japan. If momentum is large (as is often the case when the S&P500 does well - unlike Japan when the market has gone sideways, but the value premium has been large), then the momentum fund will likely outperform the S&P500 and the value funds will likely underperform with their negative momentum loads. In this respect MSCI momentum offers a great diversifier for value, as per the Asenss Journal of Finance article.

Now the GMO and RAFI forecasts are not for a 19% S&P500 return, but rather closer to zero return over the next 7-10 year (-1.9% and +0.7% real return over next 7 and 10 years respectively), with even less for US small caps (-2.9% and 0.0% real return over next 7 to 10 years respectively). Adding momentum is not likely going to save us if the actual returns are as forecast (but will help IMO). This highlights the importance for international diversification - which GMO and RAFI suggest have higher expected return over the near term (GMO's 7-year expected returns for international small are higher than international large caps +1.8% vs. +1.3% real, highest for EM at 3.5% real; RAFI's 10-year expected returns are 3.9% real for EAFE, 6.3% real for EM). Does this mean we should now sell the S&P500 and just hold EAFE and EM? A better proposition, at least for me, is diversify broadly across factors (market, size, value, and momentum; and for momentum enough to offset some of the negative momentum loads on value funds) and diversify across markets (US:EAFE:EM), and hold some high quality bonds for the 2008 type markets - then stick with it.

Merry Christmas. Best to you all.

Robert
.

garlandwhizzer
Posts: 1882
Joined: Fri Aug 06, 2010 3:42 pm

Re: Research Affiliates article on size premium

Post by garlandwhizzer » Thu Dec 25, 2014 12:19 pm

nedsaid wrote:
Investing is a very humbling experience. Once one thinks he has it all figured out, a rare event occurs in the market, and he marvels at how little he really knows. Market surprises really do happen.

I also have to allow for the fact that I might be wrong. Thus I do a whole lot of things in my portfolio. I tilt small value but I own large growth stocks. I am value oriented but I own aggressive investments. So I diversify across asset classes, diversify across factors, and frankly diversify against my own stupidity.
This sums up my own investment style and philosophy perfectly. I, too, have been humbled by the market. Anyone who invests for a long enough time will be. There are two kinds of investors who have been at it for some years. Those who have made mistakes and those who say they haven't. Diversifying widely, largely with broadly based indexes and to a lesser extent with different factors, protects me from myself.

Garland Whizzer

garlandwhizzer
Posts: 1882
Joined: Fri Aug 06, 2010 3:42 pm

Re: Research Affiliates article on size premium

Post by garlandwhizzer » Thu Dec 25, 2014 1:06 pm

Robert T wrote:
In my view a better proposition than attempting asset class rotation/timing is to set a fixed allocation, which will do okay in different market conditions - and stick with it. For example - consider the period - 1985-1999, where small value underperformed (re: the data Taylor periodically posts and reminds us of).
I believe yours is a solid investment plan, Robert T, but my religion doesn't allow me to overweight anything that I feel is fundamentally overvalued. Most of the time my approach has worked for me but not always. For example REITS were in my opinion overvalued in 2013, so I sold VNQ and moved into TSM. In 2014 VNQ massively outperformed TSM. So as Robert T says, tactical asset allocation can be tricky and difficult to execute.

Although I sometimes question factor investing on the forum, I myself do some factor investing using both momentum and value for the reasons that Robert T and others have cited. My concern is this. Now that so many funds and etfs are chasing after factors and smart beta, I suspect that factor outperformance will likely diminish or disappear in the future. The small cap value space for example occupies 3% of total market weight, small cap deep value much less, and micro cap value very small indeed. When so much money targets that small a segment it dilutes the very thing that brought the money there in the first place, value. So I believe in the logical rationale for small value but I believe that due to its increasing popularity based on factor research, it may not shine so brightly in the future. For this reason I hold only modest exposure in the range of 25% to factors and the rest to TSM in US equity.

In contrast, I believe that real fundamental value is widely available internationally and have moved to a 60/40 US/INTL split to try to take advantage of that, but like my bet on REITS, that hasn't yet paid off for me yet. I retain my faith, however, that both my losing bets, increasing INTL and decreasing REITS, based on what I believed to be fundamental valuations, will eventually win. Similar to the way that Robert T believes that the PXSV and MTUM combo will win in the long run. The difference is that my non-beta bet is limited because I'm not really sure which will outperform going forward.

Garland Whizzer

User avatar
nedsaid
Posts: 9697
Joined: Fri Nov 23, 2012 12:33 pm

Re: Research Affiliates article on size premium

Post by nedsaid » Thu Dec 25, 2014 1:30 pm

Garland, you are the perfect example of overthinking things. I am saying this because I have done exactly the same things you have. Like you, I am very concerned about valuation and think it very important as an investor. I actually have done a bit of tactical allocation but mostly with left over money market funds that were raised in 2000 when I reduced my stake in my stock mutual funds by 30%. I picked what I thought were opportune times to shift money market funds into bond funds.

About 2007, I learned about small/value tilting. One mistake that I made was that I bought International Small/Mid-Cap just before the 2008-2009 bear markets. I bought an ETF and a mutual fund. Those things got really hammered and have only recently gotten near the point at which I bought them. It was a great idea for diversification but my timing was horrible.

It was interesting that I went back through my investment records. I had recalled that I suffered about 35% losses in each of the 2000-2002 and 2008-2009 bear markets. My losses in the 2000-2002 bear were actually only about 31-32%. My tilted portfolio suffered about a 35% maybe 36% loss. I bet that if I stuck with what I had before, my losses would have looked more like what they were in 2000-2002. I think what I did was perhaps increase returns with my tilts but also volatility. According to portfolio theory, my shifts that I made in 2007-2008 (before the crash) should have also reduced volatility. But I think all I did was put more octane in the gas tank. I may have had too much "Tiger in the Tank" during the downturn.

Like you, I had big concerns about REIT valuations way, way back in 2013. Man that was so long ago I hardly remember it! :happy :happy Imaging my shock when my REIT funds went up over 25% so far this year. I trimmed a tiny bit off the top but otherwise did nothing with my REITs.

So I think Robert T's advice is really sound. Tactical asset allocation is a bit like the rebalancing bonus. Sometimes you get the bonus and sometimes you don't. The problem is that you and I are for the most part rational human beings and we expect the markets to be the same. As we have found out, markets in the short run can be very irrational. It also illustrates one of the things I learned about investing. That is if you want your investment to go up in price, sell it first. That works like a charm! :happy :happy

I am sure that if you look at your records that you will find that you made tactical allocation decisions that worked. My guess is that your track record might be 60% right and 40% wrong. Your decision to sell your REITs may yet be vindicated. If REITs were overvalued in 2013, they are even more overvalued now. But they went up and not down and that is the way the pickle squirts.
Last edited by nedsaid on Thu Dec 25, 2014 2:04 pm, edited 1 time in total.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 9697
Joined: Fri Nov 23, 2012 12:33 pm

Re: Research Affiliates article on size premium

Post by nedsaid » Thu Dec 25, 2014 2:02 pm

As Frank Sinatra said, "Regrets, I've had a few." Or as a famous poet might have said "Oh, the many investment mistakes I have made, let me count the ways." I certainly have made my share of investment mistakes and errors.

I believe that if an investor has a well thought out plan based on sound principles, that over time that investor will succeed. Maybe not with market beating results but well enough to meet life objectives. If you are a good saver, ''well enough" as an investor will do the trick. A high savings rate covers a multitude of investing errors.

The trick is to buy good stuff and to stick with it. It really helps if you can buy the good stuff at decent prices. If an investor can resist the temptation to buy at the top and sell at the bottom that will hugely increase the chances for his or her success.

I have essentially had a conservative bent as an investor. But there is a part of me that wants to continually improve and to venture out. Hence my forays into many different things. One of those being tilting my portfolio towards the smaller stocks, what we call mid-caps, small-caps, and even the micro-caps.

But Garland did a great thing with this thread and that was to challenge some assumptions. He offered some great perspectives on the size premium. Does it exist at all? The comment about large cap growth bubbles causing most of this was thought provoking. We all need from time to time to rethink our assumptions and examine them in the light of more recent data if they are still true.

What we see are differences in the way we look at the world. Our beliefs affect what we perceive about the world and that includes investments. So this is a healthy thing. A died in the wool indexer should be able to understand why some people tilt. Those of us tilt need to understand why we might be wrong. This exercise of examining our core beliefs about investments will at the same time increase what we are able to perceive about the markets. That is a good thing.
A fool and his money are good for business.

RunningRad
Posts: 324
Joined: Wed May 21, 2014 2:06 pm

Re: Research Affiliates article on size premium

Post by RunningRad » Thu Dec 25, 2014 2:13 pm

This is perhaps a bit simplistic, but it seems to me that there are two camps. One believes that small caps outperform the general market over time, and there are academic studies that demonstrate it. The other camp believes that small caps do not outperform over time, and there are academic studies that support this viewpoint.

Without further parsing the issue (separating small value from small growth, the 10th decile from the 9th, etc.), since there is no legitimate argument that small caps underperform the overall market, would not the obvious strategy be to tilt to small caps?
Few decisions in life motivated by greed ever have happy outcomes--Peter Bernstein, The 60/40 Solution

Browser
Posts: 4857
Joined: Wed Sep 05, 2012 4:54 pm

Re: Research Affiliates article on size premium

Post by Browser » Thu Dec 25, 2014 2:25 pm

Considering valuations, I find it difficult to commit to anything in the equity space these days -- but as Bogle says "we must still invest." Would be nice to run away and hide someplace. Seems to me the fuel for factor investing is basically the desperate search under every rock for some way to deal with what seems to be the horrible risk metrics for stocks. I expect that we'll all find out down the road that, however we sliced and diced, we're going to end up having meager returns to show for weathering some scary market action during the journey. Happy Trails my compadres.
We don't know where we are, or where we're going -- but we're making good time.

User avatar
JoMoney
Posts: 5258
Joined: Tue Jul 23, 2013 5:31 am

Re: Research Affiliates article on size premium

Post by JoMoney » Thu Dec 25, 2014 5:36 pm

RunningRad wrote:...Without further parsing the issue (separating small value from small growth, the 10th decile from the 9th, etc.), since there is no legitimate argument that small caps underperform the overall market, would not the obvious strategy be to tilt to small caps?
I would say the opposite. For a passive investor, small-caps don't offer a higher expected return, but they do bring risks that it isn't necessary to take. Like adding junk bonds to a high-quality bond portfolio, it's di-worse-ification.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

Browser
Posts: 4857
Joined: Wed Sep 05, 2012 4:54 pm

Re: Research Affiliates article on size premium

Post by Browser » Thu Dec 25, 2014 6:33 pm

JoMoney: Thanks for posting those Buffett quotes. Every now and then it's good to hear some plain talk. I especially liked this one, because I always thought the same thing. I hear that there are risk premiums for this and that but I have not yet determined if this is a useful concept or just an academic circularity. Now that I hear Buffett say it, I'm giving myself a bit more credibility.
Risk premiums are mostly nonsense. The world isn't calculating risk premiums.
We don't know where we are, or where we're going -- but we're making good time.

User avatar
JoMoney
Posts: 5258
Joined: Tue Jul 23, 2013 5:31 am

Re: Research Affiliates article on size premium

Post by JoMoney » Thu Dec 25, 2014 7:08 pm

Browser wrote:...I hear that there are risk premiums for this and that but I have not yet determined if this is a useful concept or just an academic circularity. Now that I hear Buffett say it, I'm giving myself a bit more credibility.
Risk premiums are mostly nonsense. The world isn't calculating risk premiums.
I think you can distill it down to your beliefs in the Efficient Market Hypothesis. If the EMH is true, the only way anyone earns excess returns is by loading up on risk (however they want to measure risk these days), and hoping the risk doesn't actually show up. If instead it's more nuanced, with elements of competitive skill and information, with behavioral factors and potential for swings in sentiment, with fads and bubbles, then I think one needs to be a bit more cautious about their approach and reluctant to except anything that would seem to promise an easy path to "above average" results.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Research Affiliates article on size premium

Post by Robert T » Fri Dec 26, 2014 2:39 am

.
Q&A with Fama on Small Value - seems relevant to overall topic
https://www.youtube.com/watch?v=HIKO-t4vU6Q

User avatar
Robert T
Posts: 2521
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Research Affiliates article on size premium

Post by Robert T » Fri Dec 26, 2014 6:40 am

.
A few thoughts on a 'Buffet approach' to investing (given some of his quotes). If we want to push aside the academic stuff and instead 'invest like Buffet' how would we go about doing it? And would it lead to superior performance?

As a benchmark, to first look back, I use my portfolio returns since inception at start of 2003 (which is based on the academic stuff), then I try to look forward (although no guarantees), comparing this to a few possible Buffet type approaches:
  • (i) Invest in S&P500 - following his advice for his wife's investments when he eventually passes on.

    (ii) Simply invest everything in Berkshire Hathaway stock (although with some risk of what happens after Warren).

    (iii) Invest in the fund he recommended to the clients of his previous partnership the Sequoia Fund (although the original principals are no longer there, and its closed to new investors - but lets assume its still open).

    (iv) Use 13F's of investors who follow the Buffet approach and just invest directly in those stocks (e.g. Mohnish Pabrai who tries to clone Buffet's approach). Another investor who follows the Buffet approach is Guy Spiers of the Aquamarine Fund (recently described his recent book, The Education of a Value Investor).

    There may be others.
First a look back to 2003 to 2014 YTD. Each of the listed portfolios matches my 2008 downside return (adding intermediate treasuries, where necessary). FWIW, my portfolio is 75:25 stock:bond, with global small cap and value tilt (same allocation, and factor exposure targets for last 12 years).
  • 10.3% = Acutal returns (based on the academic stuff)
    .8.5% = MSCI ACWI (globally divesified)
    .9.0% = S&P500
    .9.7% = Berkshire Hathaway stock
    .9.4% = Sequoia
The exact allocation to match 2008 returns were for MSCI ACWI - 75% MSCI ACWI:25% intermediate treasury; S&P500 - 83% S&P500:17% Intermediate treasury; Birkshire Hathaway stock - 93% stock:7% intermediate treasury; Sequoia = 100% Sequoia.
  • 2003-2013 (as don't have 2014 returns for the Pabrai Funds or the Aquamarine Fund)
    11.0% = Actual returns
    11.0% = Aquamarine Fund
    14.0% = Pabrai Fund
The exact allocations to match 2008 returns were for Aquamarine fund - 70% Aquamarine fund:30% Intermediate treasury. Pabrai Fund - 56% Pabrai fund: 44% Intermediate Treasury. The Pabrai fund had significant losses in 2007-08.

The 'academic stuff' is doing okay so far. Haven't missed much, at least not yet. And the 'Buffet' approach tends to result in a mid cap value and quality tilt, anyway. As per the factor loads on the Sequoia fund (for its full life), and the Aquamarine fund. Analysis on Berkshire also indicates a value/quality tilt.

Looking forward.

- It may well be the case that a small/mid value tilted portfolio loses some ground to the S&P500. Its inevitable (almost guaranteed) that there will be periods of small/mid value underperformance. The Sequoia fund lagged the S&P500 significantly from 1985-1999 - and even some investors in this fund couldn't stay the course - http://www.nytimes.com/1999/12/26/busin ... uster.html - human nature. The hardest part of investing is staying the course.

- The long-term prospects of Berkshire is uncertain - will Ted Weschler and Todd Combs deliver the goods? Hard to know, but arguably more difficult to do with such a large asset base.

- Sequoia fund is closed, and is not managed anymore by its founders Ruane (who passed on in 2005) and Cunniff (who passed on earlier this year). Will current manager Goldfarb and Poppe continue to deliver? Not sure.

- Aquamarine Fund is also closed, and based in Zurich. Pabrai funds are closed, but could still buy stocks from 13f's (although a bit delayed). Pabrai currently has a highly concentrated portfolio of about 6 stocks). His fund had significant downside in 2007-2008, not sure how it will perform going forward - but an option, perhaps not for the faint-hearted.

FWIW - this quick review doesn't seem to suggest the academic stuff is way off. Or that there is an easily accessible Buffet type approach that will deliver superior performance.

As Buffet himself says:
“To invest successfully does not require a stratospheric IQ, unusual business insights, or inside information. What’s needed is a sound intellectual framework for making decisions and the ability to keep emotions from corroding the framework.”
For me, the FF three factor model provides the intellectual framework, and has helped tremendously in making portfolio decisions. For others the framework is elsewhere, and that perfectly fine.

For those looking for a more Buffettesque ETF – the Morningstar Wide Moat Focused Index (tracked by MOAT) is interesting. It has the associated midcap value and quality tilt (based on earlier analysis by Samuel Lee http://ibd.morningstar.com/article/arti ... ,%20brf295 . Its simulated annualized returns for 2003-2014 YTD was 14.8% with 2008 downside of -19.6%. Will this performance continue? Not sure.

Many roads to Rome.

Robert
.

Browser
Posts: 4857
Joined: Wed Sep 05, 2012 4:54 pm

Re: Research Affiliates article on size premium

Post by Browser » Fri Dec 26, 2014 9:43 am

Very nice and helpful post, Robert. I think the Buffett quote about having a sound rationale which provides the foundation for staying the course is key. Your quote:
For me, the FF three factor model provides the intellectual framework, and has helped tremendously in making portfolio decisions. For others the framework is elsewhere, and that perfectly fine.
Quite a reasonable position. Even though I love the intellectual, academic approach to doing everything, my biggest problem with this approach in managing my investment portfolio is believing that I've figured out the specifics of actually implementing it in the real world. For example, even if I really believe the academic story about momentum, I find it impossible to believe strongly enough that MTUM is how to actualize it. There's a big difference between believing in the theory and believing in the funds and ETFs I've got real money invested in, and there's the rub.

One thing I do think is true. The best ways to make money are to do things like start a business, become a landlord and manage property, devote our efforts toward career advancement, etc. All those things involve doing a lot of difficult and often unpleasant work 24/7. Those of us who prefer to sit in front of our computers and armchair our capital should be realistic about our wealth expectations. You might become wealthy doing that - but at the end of the day it's probably because you got lucky and not because of the hard work you did. You know: the old story about the hikers who got hopelessly lost and found their way to safety by inadvertently using the wrong map. Most of us would probably be better off not overthinking this stuff. It's probably more important to believe that you have a good map and religiously follow it than to actually have one. Eventually, if capital markets continue to work and we don't destroy the planet, we'll all probably wind up in about the same place.
We don't know where we are, or where we're going -- but we're making good time.

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

Re: Research Affiliates article on size premium

Post by richard » Fri Dec 26, 2014 10:33 am

Robert T wrote:<>For me, the FF three factor model provides the intellectual framework, and has helped tremendously in making portfolio decisions. For others the framework is elsewhere, and that perfectly fine.<>
The interpretation of the three factor model for investment decisions is far from clear. The two most popular are:
- SV represent risks. If you want more risk in the hope of more return, tilt to SV (and vice versa). Alas, risk means return may be lower, perhaps substantially
- SV historical performance is due to behavioral anomalies that are likely to persist. Close to a free lunch, although patience is required.

Do you agree with either of these? If so, which? If not, what's your interpretation?

garlandwhizzer
Posts: 1882
Joined: Fri Aug 06, 2010 3:42 pm

Re: Research Affiliates article on size premium

Post by garlandwhizzer » Fri Dec 26, 2014 2:29 pm

Robert T's suggestion to look at MOAT as an example of Buffett's approach is a good and interesting one. The rationale behind the etf is compellingly presented in their video presentation available at http://www.vaneck.com. Its approach winds up with a tilt toward value in LC and a tilt toward growth in MC at least according to Morningstar's style boxes. It clearly attempts to replicate Buffett's methodology and has performed well during its existence since 4/12. It comes an expense ratio of .49%, so it has to outperform comparable indexes by a least a half percent annually to actually make money for investors. It very clearly makes highly concentrated bets on a few companies and only has 21 stocks in its portfolio, a potential downside in terms of diversification. Apparently they feel that their wide moat companies have sufficient quality and intrinsic value that they don't need diversification to offset concentrated portfolio risk.

Garland Whizzer

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

Re: Research Affiliates article on size premium

Post by hafius500 » Sat Dec 27, 2014 3:46 pm

JoMonbey wrote:I think you can distill it down to your beliefs in the Efficient Market Hypothesis. If the EMH is true, the only way anyone earns excess returns is by loading up on risk (however they want to measure risk these days), and hoping the risk doesn't actually show up. If instead it's more nuanced, with elements of competitive skill and information, with behavioral factors and potential for swings in sentiment, with fads and bubbles, then I think one needs to be a bit more cautious about their approach and reluctant to except anything that would seem to promise an easy path to "above average" results.
But does the EMH really claim that a higher expected return is the reward for taking higher 'risks'?
Does the EMH rest upon a specific asset-pricing model that relates expected returns to 'risks'?

IIRC, Samuelson's paper on the impossibility of efficient markets concluded that the price-setting investors (i.e.,those who collect and efficiently process the information) should earn a "premium".
In such a model superior information and knowledge would be linked to a higher expected return and it would be consistent with the practitioner's definition of risk:
Risk comes from not knowing what you are doing.

Isn't this definition consistent with the EMH?

Didn't Illmanen ("Expected Returns") write that bonds with (high) default risks don't necessarily have to offer higher expected returns than bonds without default risks ?

It has been written that economics is a system of moral values disguised as science.

What are the (moral? theological?) roots of the risk-reward theory? I couldn't find any Wikipedia articles about the sources of financial and economic returns (besides articles about models which are tautological).

In the real economy, the proverbial and successful "prudent merchant" intends to minimize 'risks'. He is not a risk-taker.

In modern financial theories, Forrest Gump can expect to earn a higher profit than experts.
prior username: hafis50

Post Reply