Small Cap Value worth it?
Small Cap Value worth it?
I'm thinking about tilting my stock allocation towards small cap value. I had a quick look at the performance of the Vanguard Small Cap Value ETF (VBR) and the performance compared with the US Total Stock Market ETF (VTI) seems very similar:
1 Year return: VBR 4% vs VTI 7%
5 Year return: 82% vs 79%
10 year return: 82% vs 79% (no typo, 5 and 10 year returns are exactly the same).
Returns are price only, excluding dividends.
Given that the performance has been very similar, is tilting worth it? And if so, how much would you tilt?
Also curious if people tilt for international stocks and if so, what Vanguard Fund / ETF do you use?
thanks
1 Year return: VBR 4% vs VTI 7%
5 Year return: 82% vs 79%
10 year return: 82% vs 79% (no typo, 5 and 10 year returns are exactly the same).
Returns are price only, excluding dividends.
Given that the performance has been very similar, is tilting worth it? And if so, how much would you tilt?
Also curious if people tilt for international stocks and if so, what Vanguard Fund / ETF do you use?
thanks
Re: Small Cap Value worth it?
My asset allocation is 30% stocks/70% bonds.
Within that, I tilt 70% domestic stock, 30% international
Within that, I tilt 50% TSM/50% small cap value. With international it's a 50/50 split 'twixt large and small cap.
This is, more or less, a combination of Larry Swedroe's "eliminating fat tails" allocation and Trev H's "ultimate buy and hold" equities allocation. (I don't tilt value in international, so it's a bit different.)
Will this allocation make a big improvement over a Three-Fund Portfolio going forward? I'm willing to base some investing decisions on the proposition that it will make some difference, but since I handed in my amazing psychic powers when I graduated high school (Mom insisted), I could be wrong .
I also don't think it will make a huge difference one way or the other. I'm realistic, and have been at this long enough to know that the 30/70 allocation (up at the top) is the most important allocation I've listed.
Within that, I tilt 70% domestic stock, 30% international
Within that, I tilt 50% TSM/50% small cap value. With international it's a 50/50 split 'twixt large and small cap.
This is, more or less, a combination of Larry Swedroe's "eliminating fat tails" allocation and Trev H's "ultimate buy and hold" equities allocation. (I don't tilt value in international, so it's a bit different.)
Will this allocation make a big improvement over a Three-Fund Portfolio going forward? I'm willing to base some investing decisions on the proposition that it will make some difference, but since I handed in my amazing psychic powers when I graduated high school (Mom insisted), I could be wrong .
I also don't think it will make a huge difference one way or the other. I'm realistic, and have been at this long enough to know that the 30/70 allocation (up at the top) is the most important allocation I've listed.
-
- Posts: 41
- Joined: Sun Dec 28, 2014 1:27 pm
Re: Small Cap Value worth it?
I use VSS for international small cap and EFV for international value. I'm unaware of any small cap value int'l index funds outside of DFA...
-
- Posts: 1726
- Joined: Fri Jan 03, 2014 7:31 pm
- Location: Kansas City, MO
Re: Small Cap Value worth it?
I tilt small cap in US/non-US equity.
Core4 + 2 small cap's.
10% - VSS - Vanguard FTSE All World ex US Small Cap
10% - IJS - iShares S&P Small Cap 600 Value
I'm thinking about selling the IJS for VBR as everything else in my Fidelity port is now Vanguard (go figure) but the extra REIT exposure is troubling me. I might even start looking around for a better ETF than VSS but I will stick with the 10% tilt in both sectors.
Core4 + 2 small cap's.
10% - VSS - Vanguard FTSE All World ex US Small Cap
10% - IJS - iShares S&P Small Cap 600 Value
I'm thinking about selling the IJS for VBR as everything else in my Fidelity port is now Vanguard (go figure) but the extra REIT exposure is troubling me. I might even start looking around for a better ETF than VSS but I will stick with the 10% tilt in both sectors.
70% AVGE | 20% FXNAX | 10% T-Bill/Muni
Re: Small Cap Value worth it?
If you are going to buy once and hold, then it probably doesn't make sense to tilt.
I think the bonus comes in when you buy low and sell high during rebalancing moves. I posted these numbers in another thread, but here are the 2014 total return numbers for 4 funds in my 401(k) and my returns in those funds because of market timing rebalancing moves:
LS vs B&H
14.12 13.62 FUSVX S&P500 index fund
11.91 07.71 FSEVX Extended market index fund
-4.36 -4.40 FGSDX Total Intl index fund
06.02 05.94 FSITX Total US Bond Index fund
In every instance the LS return is higher. This is because small-caps were more volatile and dropped more making it relatively easy to decide when to rebalance into them.
FWIW, I tilt heavily to small-caps. In a 9-box M* style grid I usually have about 45% large-caps and 55% mid/small-caps for both US and foreign equities. I've posted about that in this thread: http://www.bogleheads.org/forum/viewtop ... &p=2248842
I think the bonus comes in when you buy low and sell high during rebalancing moves. I posted these numbers in another thread, but here are the 2014 total return numbers for 4 funds in my 401(k) and my returns in those funds because of market timing rebalancing moves:
LS vs B&H
14.12 13.62 FUSVX S&P500 index fund
11.91 07.71 FSEVX Extended market index fund
-4.36 -4.40 FGSDX Total Intl index fund
06.02 05.94 FSITX Total US Bond Index fund
In every instance the LS return is higher. This is because small-caps were more volatile and dropped more making it relatively easy to decide when to rebalance into them.
FWIW, I tilt heavily to small-caps. In a 9-box M* style grid I usually have about 45% large-caps and 55% mid/small-caps for both US and foreign equities. I've posted about that in this thread: http://www.bogleheads.org/forum/viewtop ... &p=2248842
-
- Posts: 1726
- Joined: Fri Jan 03, 2014 7:31 pm
- Location: Kansas City, MO
Re: Small Cap Value worth it?
I agree with livesoft, I tend to fiddle once a year on my birthday with new money/re-balance money.
It's better than when I was buying/selling stocks pre-2008 but it's not buy once and forget.
If I was that type I would go Core4 and forget about tilting altogether as it's not needed if on a re-balance only once a decade plan.
It's better than when I was buying/selling stocks pre-2008 but it's not buy once and forget.
If I was that type I would go Core4 and forget about tilting altogether as it's not needed if on a re-balance only once a decade plan.
70% AVGE | 20% FXNAX | 10% T-Bill/Muni
Re: Small Cap Value worth it?
I'm also wondering where you got your return numbers. VBR returned 10.55% in 2014 with a $1.869 distribution, so around a 2% distribution to be generous. You stated 4%. Maybe you are thinking of some other VBR? Full disclosure: I own lots of VBR.
Re: Small Cap Value worth it?
Thanks for the feedback. Return numbers are as per today as per Google Finance.
Re: Small Cap Value worth it?
livesoft, by "rebalancing moves" do you mean rebalancing couple of times a year or something else?
edited to fix typos.
edited to fix typos.
Re: Small Cap Value worth it?
I mean things like this: viewtopic.php?f=10&t=153585
Re: Small Cap Value worth it?
OP provided returns excluding dividends, NAV only.
Before you consider tilting your asset allocation, please educate yourself on why you always need to consider dividends. It is a pretty basic concept you should understand before fiddling with your portfolio like this.
Before you consider tilting your asset allocation, please educate yourself on why you always need to consider dividends. It is a pretty basic concept you should understand before fiddling with your portfolio like this.
Re: Small Cap Value worth it?
I get the concept, that's why I clearly stated that the returns were prices only excluding dividends.
Re: Small Cap Value worth it?
Performance was similar - but how about correlation?ge1 wrote:I'm thinking about tilting my stock allocation towards small cap value. I had a quick look at the performance of the Vanguard Small Cap Value ETF (VBR) and the performance compared with the US Total Stock Market ETF (VTI) seems very similar:
1 Year return: VBR 4% vs VTI 7%
5 Year return: 82% vs 79%
10 year return: 82% vs 79% (no typo, 5 and 10 year returns are exactly the same).
Returns are price only, excluding dividends.
Given that the performance has been very similar, is tilting worth it? And if so, how much would you tilt?
Also curious if people tilt for international stocks and if so, what Vanguard Fund / ETF do you use?
thanks
Leonard |
|
Market Timing: Do you seriously think you can predict the future? What else do the voices tell you? |
|
If employees weren't taking jobs with bad 401k's, bad 401k's wouldn't exist.
- sunnywindy
- Posts: 655
- Joined: Sat Jan 18, 2014 3:42 pm
- Location: Central California
Re: Small Cap Value worth it?
Here is a Morningstar chart of Vanguard Small Cap (VB), Small Cap Value (VBR), and Small Cap Growth (VBK) since inception. Clearly, Growth has won the day since the market bottom. Precrash they are much more even. Personally, I overweight small caps, but do not tilt value or growth.
http://tinyurl.com/nhf82hk
http://tinyurl.com/nhf82hk
Powered by chocolate!
Re: Small Cap Value worth it?
Then you should know that information is basically useless. Why not use a source that provides the actual returns?ge1 wrote:I get the concept, that's why I clearly stated that the returns were prices only excluding dividends.
This is also only a price chart.sunnywindy wrote:Here is a Morningstar chart of Vanguard Small Cap (VB), Small Cap Value (VBR), and Small Cap Growth (VBK) since inception. Clearly, Growth has won the day since the market bottom. Precrash they are much more even. Personally, I overweight small caps, but do not tilt value or growth.
http://tinyurl.com/nhf82hk
-
- Posts: 232
- Joined: Wed May 14, 2008 3:48 am
Re: Small Cap Value worth it?
I tilt toward SCV about 5x market cap or about ~15% of my domestic stock allocation in my retirement accounts.
I also tilt my kids' 529 plans with 25% of the total allocation to SCV with a DFA fund in the Utah plan (the rest in is total stock and total international indexes).
I also tilt toward small cap international with Schwab's ETF SCHC about 2x market cap or 22%.
I also tilt my kids' 529 plans with 25% of the total allocation to SCV with a DFA fund in the Utah plan (the rest in is total stock and total international indexes).
I also tilt toward small cap international with Schwab's ETF SCHC about 2x market cap or 22%.
Re: Small Cap Value worth it?
I tilt 50% TSM/50% SCV as some of the others here. I do this to get exposure to the HmL and SmB factors and for the small rebalacing benefit. I swallow the pill and use VISVX for this, although would go DFA if I could. My 1 year old daughter gets the DFA funds through her Utah 529. She has a Cadillac tilted portfolio and I have the Fidelity Toyota Corolla equivalent in my 401(k)!
My IPS is informed by the academic research in factor investing, so I SWAN with this large of a tilt. If it worked or not, we'll see in 30 years. I would bet that the fact that I am a young supersaver will probably matter a ton more than my allocation.
My IPS is informed by the academic research in factor investing, so I SWAN with this large of a tilt. If it worked or not, we'll see in 30 years. I would bet that the fact that I am a young supersaver will probably matter a ton more than my allocation.
Re: Small Cap Value worth it?
How confident are we that the value premium will appear in a standard investment lifetime of 40 years or so - and that hedging growth and value with small cap blend would not have been the best choice in small cap land? Vanguard's small cap growth, annualized the last 10 years has returned 9.6% versus small cap value's 8.3%. So far I'm glad I hedge and I do realize this is a short sighted observation.
I've always had this dilemma with anything other than a size factor; the small tilt to me is very concrete and logical, the value tilt seems more abstract. How do the computers even flag "value" stocks when many times "E" of (P/E) isn't even a real number (it includes a bunch of one time expenses, non GAAP stuff, etc). Do the computers use the adjusted "E" like everyone else on Wall Street does? Do the computers use other statistics besides P/E? OK I've looked and partially answered my question:
CRSP classifies value securities using the following factors: book to price, forward earnings to price, historic earnings to price, dividend-to-price ratio and sales-to-price ratio.
I still remain more confident in small than value. I know it is irrational. I'm off to search for old conversation threads on the matter.
I've always had this dilemma with anything other than a size factor; the small tilt to me is very concrete and logical, the value tilt seems more abstract. How do the computers even flag "value" stocks when many times "E" of (P/E) isn't even a real number (it includes a bunch of one time expenses, non GAAP stuff, etc). Do the computers use the adjusted "E" like everyone else on Wall Street does? Do the computers use other statistics besides P/E? OK I've looked and partially answered my question:
CRSP classifies value securities using the following factors: book to price, forward earnings to price, historic earnings to price, dividend-to-price ratio and sales-to-price ratio.
I still remain more confident in small than value. I know it is irrational. I'm off to search for old conversation threads on the matter.
-
- Posts: 1726
- Joined: Fri Jan 03, 2014 7:31 pm
- Location: Kansas City, MO
Re: Small Cap Value worth it?
Just curious what would a "Cadillac tilted DFA portfolio" look like?tarheel wrote:I tilt 50% TSM/50% SCV as some of the others here. I do this to get exposure to the HmL and SmB factors and for the small rebalacing benefit. I swallow the pill and use VISVX for this, although would go DFA if I could. My 1 year old daughter gets the DFA funds through her Utah 529. She has a Cadillac tilted portfolio and I have the Fidelity Toyota Corolla equivalent in my 401(k)!
My IPS is informed by the academic research in factor investing, so I SWAN with this large of a tilt. If it worked or not, we'll see in 30 years. I would bet that the fact that I am a young supersaver will probably matter a ton more than my allocation.
I've never looked at anything DFA.
70% AVGE | 20% FXNAX | 10% T-Bill/Muni
Re: Small Cap Value worth it?
Unscientific/?meaningless observations of up to 15 year Morningstar returns:
1) Marginal difference between Total Market and S&P 500: 5 to 15 years, former ahead by 0.26 to 0.61%
2) Small cap blend specifically Tax Managed Small Cap has beaten Total: 5 to 15 years 1.66 to 5.05%.
SCV and SCG have had variable returns with the former winning only at the 15 year mark by 6.5%.
Dimensional SCV has only beaten Total Market at 5 and 15 years by 0.39 and 6.5.
3) Total Stock outdid SCV in 2007, 2011 and 2013.
Take away points:
1) Small caps are a separate subasset class which offer diversification, rebalancing and potential for higher return.
See also http://www.servowealth.com/resources/ar ... sset-class
2) Return comparisons are highly period dependent and are a moving target.
3) Any style smallcap has been good but SV should be default in view of strong academic support and outperformance over longer periods.
You have to be prepared to wait for the payoff though you may pick up some points with rebalancing on the way.
4) Vanguard SG ain't no black hole.
5) Vanguard Tax Managed Small Cap has held up very well and I hold it too!
6) If you are into market timing, recent underperformance of SV should actually be a buy signal! Long termers should be rebalancing, of course.
Venkat
1) Marginal difference between Total Market and S&P 500: 5 to 15 years, former ahead by 0.26 to 0.61%
2) Small cap blend specifically Tax Managed Small Cap has beaten Total: 5 to 15 years 1.66 to 5.05%.
SCV and SCG have had variable returns with the former winning only at the 15 year mark by 6.5%.
Dimensional SCV has only beaten Total Market at 5 and 15 years by 0.39 and 6.5.
3) Total Stock outdid SCV in 2007, 2011 and 2013.
Take away points:
1) Small caps are a separate subasset class which offer diversification, rebalancing and potential for higher return.
See also http://www.servowealth.com/resources/ar ... sset-class
2) Return comparisons are highly period dependent and are a moving target.
3) Any style smallcap has been good but SV should be default in view of strong academic support and outperformance over longer periods.
You have to be prepared to wait for the payoff though you may pick up some points with rebalancing on the way.
4) Vanguard SG ain't no black hole.
5) Vanguard Tax Managed Small Cap has held up very well and I hold it too!
6) If you are into market timing, recent underperformance of SV should actually be a buy signal! Long termers should be rebalancing, of course.
Venkat
-
- Posts: 41
- Joined: Sun Dec 28, 2014 1:27 pm
Re: Small Cap Value worth it?
But then most say momentum is the better way to go, not the recent under-performers?vencat wrote:6) If you are into market timing, recent underperformance of SV should actually be a buy signal! Long termers should be rebalancing, of course.
Re: Small Cap Value worth it?
You say you "get the concept", in which case you must be aware that comparing price without dividends is meaningless.ge1 wrote:I get the concept, that's why I clearly stated that the returns were prices only excluding dividends.
Re: Small Cap Value worth it?
You need both, at least according to this articleBabylonOrBust wrote:But then most say momentum is the better way to go, not the recent under-performers?vencat wrote:6) If you are into market timing, recent underperformance of SV should actually be a buy signal! Long termers should be rebalancing, of course.
http://www.investingdaily.com/16052/the ... -momentum/
Personally, I don't have any momentum exposure except through VTSMX.
Re: Small Cap Value worth it?
PDN, IEIS, SFILX, FNDC - Dev SCV tilted (I use the last 2)BabylonOrBust wrote:I use VSS for international small cap and EFV for international value. I'm unaware of any small cap value int'l index funds outside of DFA...
DGS - EM SCV tilted (I use)
For US I use PXSV in IRA and VIOV/IJS & VBR in taxable
-
- Posts: 41
- Joined: Sun Dec 28, 2014 1:27 pm
Re: Small Cap Value worth it?
Thanks for the info. Of the first group, only IEIS really differs from VSS in allocation according to Morningstar. IEIS looks really good though.caklim00 wrote:PDN, IEIS, SFILX, FNDC - Dev SCV tilted (I use the last 2)BabylonOrBust wrote:I use VSS for international small cap and EFV for international value. I'm unaware of any small cap value int'l index funds outside of DFA...
DGS - EM SCV tilted (I use)
For US I use PXSV in IRA and VIOV/IJS & VBR in taxable
- nisiprius
- Advisory Board
- Posts: 52211
- 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: Small Cap Value worth it?
First, before I show the charts, some explanations... I think I'm using MPT methodology correctly. I don't want to go down the rathole of whether this methodology is useful or silly, but it is part of the rationale behind "diversification benefit" arguments. The data for the first chart comes from the "Simba backtest spreadsheet" maintained by the forum. The Fama-French data for the second chart is from the Kenneth French data library. The charts were created with my own program and might be buggy. (I have several "doh!" emoticons cued up and ready to go).
The results will be ambiguous and opinions will vary. My own opinion is that the effect is so tenuous and dubious that I am not convinced it's worth going after, and I don't, personally.
Here's the methodology. We start with the actual historic data, in this case for the 29-year period 1985 through 2013. We ask this specific question: suppose we knew that this coming year, the returns of two assets will be the same as they were in of one randomly selected past year from 1985 through 2013. Suppose we wanted to explore mixes of three assets: two mutual funds and cash. And suppose we wanted a specific amount of "risk," specifically a standard deviation of 10, which I chose because it's a moderate risk that's close to the history standard deviation of a 60/40 stock/bond portfolio.
a) What is the expected standard deviation ("risk") and return for mixes of the two funds?
b) How much improvement can we get in our return, for the same specified risk, by using a mix of the two funds rather than simply going 100% with the fund that has the higher risk-adjusted return?
By "cash," I mean "riskless asset," where riskless doesn't mean free from black swans, but means stable and free from market fluctuations. I used the average return of 3-month Treasury bills over the period 1985 through 2013.
First, two real-life mutual funds, Vanguard Total Stock Market Index and Vanguard Small-Cap Value Index.
The risk and return are almost the same. Furthermore, the red dot (Total Stock) is almost on, slightly below the green line. The fact that they're almost on the line shows that their Sharpe ratio, the risk-adjusted reward, is almost the same. Small-cap value has higher return and higher risk, but only very slight more risk-adjusted return.
The curve represents all of the (long-only) mixes of the two mutual funds, and the dots are at 10% increments--that is, the red dot is pure TSM, the first block dot is 90% TSM, 10% VISVX, the next is 80%/20%, and so on.
The place where the yellow tangent line, the "capital market line," grazes the curve tells us the optimum mix, and the angle between the yellow line and the green line tells us how much improvement we get over just using 100% VISVX. If we decide that we want our total portfolio of VTSMX, VISVX, and cash to have a risk represented by a standard deviation of 10, then using just cash and small-cap value gives us a 9.03% return, while using a mix of 45% VTSMX, 55% VISVX gives us a 9.30% return, for an improvement of 0.28%.
This is a good illustration of the supposed diversification benefit of low correlation. The correlation, ρ, is 0.74. The curve bends. Because of that bend, we have the "free lunch" in which the risk-adjusted-reward of the 45/55 mix is slightly better than that of either VTSMX or VISVX itself.
But, it isn't very much. It is only an improvement of 0.28%. And keep in mind that it's not even close to certain that we'll get it in real life.
Now, because some people think Vanguard's small value fund is icky poo and isn't small enough and "valuey" enough, I redid this and replaced the Vanguard fund results with the actual Fama-French "S/H" research portfolio, which I assume we can take to be absolutely 100% pure small value, purer than anything we can buy in any real-world mutual fund:
Now, it is absolutely true that the Fama-French theoretical data have much higher return than the Vanguard fund, but also (surprise?) much higher risk. The general pattern is much the same as before. VTSMX and small-cap value both lie almost exactly on the same capital market line and have almost the same risk-adjusted reward (Sharpe ratio). Small-cap value does have slightly higher risk-adjusted reward than VTSMX, but not much. Once again, the correlation is about the same, 0.696, the curve does bend but not very much, the optimum mix is still close to 50/50--in this case, 55 VTSMX, 45 pure small value. And once again, the improvement is there but not very large--about 0.36% this time.
Here's my point. On the face of it, using this rather theoretical framework--all of which is fragile in the real world--the "diversification" benefit, the benefit of mixing two assets instead of just going all-in with the "better" asset--is there, but it's small. And the reason is that even though an 0.7 correlation isn't a perfect 1.0, it doesn't buy you very much. Neither asset is a very powerful diversifier for the other.
Now, the what-if. Suppose that instead of a correlation of 0.7, the two assets had zero correlation. This is NOT found in the real world between any two subclasses of stocks. If we take the same means and standard deviations as VTSMX and Fama-French small value, but assume that the two assets had zero correlation, the chart would look like this:
Because this time the correlation really is low, genuinely low, not just "not 1.00," the curve bends a lot, the capital market line is pushed up a lot, and the improvement in return is not just 20 or 30 basis points, but an improvement from 9.01% to 11.03%, a gain of 2.02%.
My summary: it is true that with imperfectly correlated assets, a mix can give you more return for the same risk than either of the two assets by itself, BUT, when the correlation is in the 0.70-0.75 range that improvement is awfully small, and not remotely comparable to the improvement from truly low correlation.
And thus, my conclusion is that if we are talking about a diversification benefit from holding both Total Stock and small-cap value, the diversification benefit is awfully small. If small-cap value had near-zero correlation with the total stock market, I would become a convert and tilt.
The results will be ambiguous and opinions will vary. My own opinion is that the effect is so tenuous and dubious that I am not convinced it's worth going after, and I don't, personally.
Here's the methodology. We start with the actual historic data, in this case for the 29-year period 1985 through 2013. We ask this specific question: suppose we knew that this coming year, the returns of two assets will be the same as they were in of one randomly selected past year from 1985 through 2013. Suppose we wanted to explore mixes of three assets: two mutual funds and cash. And suppose we wanted a specific amount of "risk," specifically a standard deviation of 10, which I chose because it's a moderate risk that's close to the history standard deviation of a 60/40 stock/bond portfolio.
a) What is the expected standard deviation ("risk") and return for mixes of the two funds?
b) How much improvement can we get in our return, for the same specified risk, by using a mix of the two funds rather than simply going 100% with the fund that has the higher risk-adjusted return?
By "cash," I mean "riskless asset," where riskless doesn't mean free from black swans, but means stable and free from market fluctuations. I used the average return of 3-month Treasury bills over the period 1985 through 2013.
First, two real-life mutual funds, Vanguard Total Stock Market Index and Vanguard Small-Cap Value Index.
The risk and return are almost the same. Furthermore, the red dot (Total Stock) is almost on, slightly below the green line. The fact that they're almost on the line shows that their Sharpe ratio, the risk-adjusted reward, is almost the same. Small-cap value has higher return and higher risk, but only very slight more risk-adjusted return.
The curve represents all of the (long-only) mixes of the two mutual funds, and the dots are at 10% increments--that is, the red dot is pure TSM, the first block dot is 90% TSM, 10% VISVX, the next is 80%/20%, and so on.
The place where the yellow tangent line, the "capital market line," grazes the curve tells us the optimum mix, and the angle between the yellow line and the green line tells us how much improvement we get over just using 100% VISVX. If we decide that we want our total portfolio of VTSMX, VISVX, and cash to have a risk represented by a standard deviation of 10, then using just cash and small-cap value gives us a 9.03% return, while using a mix of 45% VTSMX, 55% VISVX gives us a 9.30% return, for an improvement of 0.28%.
This is a good illustration of the supposed diversification benefit of low correlation. The correlation, ρ, is 0.74. The curve bends. Because of that bend, we have the "free lunch" in which the risk-adjusted-reward of the 45/55 mix is slightly better than that of either VTSMX or VISVX itself.
But, it isn't very much. It is only an improvement of 0.28%. And keep in mind that it's not even close to certain that we'll get it in real life.
Now, because some people think Vanguard's small value fund is icky poo and isn't small enough and "valuey" enough, I redid this and replaced the Vanguard fund results with the actual Fama-French "S/H" research portfolio, which I assume we can take to be absolutely 100% pure small value, purer than anything we can buy in any real-world mutual fund:
Now, it is absolutely true that the Fama-French theoretical data have much higher return than the Vanguard fund, but also (surprise?) much higher risk. The general pattern is much the same as before. VTSMX and small-cap value both lie almost exactly on the same capital market line and have almost the same risk-adjusted reward (Sharpe ratio). Small-cap value does have slightly higher risk-adjusted reward than VTSMX, but not much. Once again, the correlation is about the same, 0.696, the curve does bend but not very much, the optimum mix is still close to 50/50--in this case, 55 VTSMX, 45 pure small value. And once again, the improvement is there but not very large--about 0.36% this time.
Here's my point. On the face of it, using this rather theoretical framework--all of which is fragile in the real world--the "diversification" benefit, the benefit of mixing two assets instead of just going all-in with the "better" asset--is there, but it's small. And the reason is that even though an 0.7 correlation isn't a perfect 1.0, it doesn't buy you very much. Neither asset is a very powerful diversifier for the other.
Now, the what-if. Suppose that instead of a correlation of 0.7, the two assets had zero correlation. This is NOT found in the real world between any two subclasses of stocks. If we take the same means and standard deviations as VTSMX and Fama-French small value, but assume that the two assets had zero correlation, the chart would look like this:
Because this time the correlation really is low, genuinely low, not just "not 1.00," the curve bends a lot, the capital market line is pushed up a lot, and the improvement in return is not just 20 or 30 basis points, but an improvement from 9.01% to 11.03%, a gain of 2.02%.
My summary: it is true that with imperfectly correlated assets, a mix can give you more return for the same risk than either of the two assets by itself, BUT, when the correlation is in the 0.70-0.75 range that improvement is awfully small, and not remotely comparable to the improvement from truly low correlation.
And thus, my conclusion is that if we are talking about a diversification benefit from holding both Total Stock and small-cap value, the diversification benefit is awfully small. If small-cap value had near-zero correlation with the total stock market, I would become a convert and tilt.
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.
- saltycaper
- Posts: 2650
- Joined: Thu Apr 24, 2014 8:47 pm
- Location: The Tower
Re: Small Cap Value worth it?
Just to put the seemingly minor return difference in perspective, 9.02 vs 9.30 on just $100K over 30 years should add up to something like a $100K difference in final portfolio value. Nothing to sneeze at. (A little less if these returns do not include the slightly larger ER for small cap value).
The only question is whether the difference will show up, which of course we cannot be sure of either way.
The only question is whether the difference will show up, which of course we cannot be sure of either way.
Quod vitae sectabor iter?
- nisiprius
- Advisory Board
- Posts: 52211
- 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: Small Cap Value worth it?
And there is always the tyranny of endpoints. All of the MPT parameters are every bit as capricious and fickle as return itself. Here is a 15-year period over which 100% total market was the optimum, and any admixture of small-cap value into VTSMX simply made things worse by adding risk without adding any return (reducing it, even):
Notice that in the first chart, despite an 0.67 imperfect correlation there is no diversification benefit at all. In order to get any diversification benefit with this set of parameters, the correlation needs to be lower than 0.5, but even at 0.25 the improvement is only 0.19%. At zero correlation if you could get it, the improvement would be 0.70%.
And here is one in which 100% small-cap value was the optimum, and any admixture of total market simply reduced return without reducing risk:
Again, in the second chart, despite an 0.79 correlation there was no diversification benefit at all.
Notice that in the first chart, despite an 0.67 imperfect correlation there is no diversification benefit at all. In order to get any diversification benefit with this set of parameters, the correlation needs to be lower than 0.5, but even at 0.25 the improvement is only 0.19%. At zero correlation if you could get it, the improvement would be 0.70%.
And here is one in which 100% small-cap value was the optimum, and any admixture of total market simply reduced return without reducing risk:
Again, in the second chart, despite an 0.79 correlation there was no diversification benefit at all.
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.
Re: Small Cap Value worth it?
Could your chart methodology figure out when was the optimal time to do rebalancing? Perhaps when the returns of small-cap value and total stock market diverged "significantly" and not on a calendar date?
- nisiprius
- Advisory Board
- Posts: 52211
- 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: Small Cap Value worth it?
No.livesoft wrote:Could your chart methodology figure out when was the optimal time to do rebalancing? Perhaps when the returns of small-cap value and total stock market diverged "significantly" and not on a calendar date?
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.
Re: Small Cap Value worth it?
Thanks. It looks like a nice project for me in the future.
Re: Small Cap Value worth it?
Great analysis nisiprius!
My equity side is in total us stock, scv, total intl, and ftse small exUs.
You have me considering selling my 20% equity stake of scv and moving it equally to vtsax and vtiax. (Total us stock and total intl).
Would you be able to do a check on diversification benefit chart for total us stock and total international, and possibly total international and ftse small ex us?
Sincerely
Jainn
My equity side is in total us stock, scv, total intl, and ftse small exUs.
You have me considering selling my 20% equity stake of scv and moving it equally to vtsax and vtiax. (Total us stock and total intl).
Would you be able to do a check on diversification benefit chart for total us stock and total international, and possibly total international and ftse small ex us?
Sincerely
Jainn
- nisiprius
- Advisory Board
- Posts: 52211
- 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: Small Cap Value worth it?
Jainn, YIKES! Please do not change your portfolio on the basis of anything I've posted above. This is just casual exploration. Just because I wrote my own computer program and made some pretty pictures does not mean I have a profound understanding. Don't put unwarranted trust in anyone's pretty pictures, including mine.
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.
-
- Posts: 41
- Joined: Sun Dec 28, 2014 1:27 pm
Re: Small Cap Value worth it?
I'm conflicted on it all. Nisiprius, what was the reasoning for not including periods prior to 1985?