Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 10:31 am

An argument made on this board is that factor tilts increase diversification. Factor tilted portfolios have exposure to the market factor (beta), but also exposure to other factors. Two popular factors here are small and value.

VTI (Vanguard Total Stock Market ETF) has 3641 stocks AVerage market cap is about $8.232 billion. Total market cap is about 29,972 billion.

As an example of a factor tilted fund, IJS (iShares S&) Small-Cap 600 Value ETF) has 462 stocks. Average market cap is about $1.24011 billion. Total market cap is about $573 billion.

IJS has 12.7% of the number of stocks that VTI has. If you assume that VTI represents 100% of the US investable stock market, then IJS is about 1.91% of the total US stock market by market cap.

IJS is much more concentrated than VTI. Aren't you taking on considerable idiosyncratic risk with factor tilted funds, such as IJS?

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by DaufuskieNate » Sun Dec 16, 2018 10:48 am

Just another perspective: The top 10 holdings of VTI represent 17.7% of the total. The top 10 for IJS represent 7.6%.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 11:11 am

VTI is concentrated in megacap/large cap stocks.

The following is from someone else, not me:

"As for the weaknesses of indexes: the megacap issue doesn't give me that much pause, because I view the megacaps as inherently more diversified; e.g. should you underweight Exxon vs a collection of small caps with the same number of oil fields in the aggregate?...Even something as concentrated as Apple has moving pieces, the iPhone markets often move very differently from one another"

Random Walker
Posts: 4233
Joined: Fri Feb 23, 2007 8:21 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Random Walker » Sun Dec 16, 2018 11:20 am

I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave

stan1
Posts: 8023
Joined: Mon Oct 08, 2007 4:35 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by stan1 » Sun Dec 16, 2018 11:29 am

Park wrote:
Sun Dec 16, 2018 10:31 am
IJS is much more concentrated than VTI. Aren't you taking on considerable idiosyncratic risk with factor tilted funds, such as IJS?
Most people who choose to invest in Small Value do so as a tilt, not as a replacement for a total market fund. Any single company is a very, very small contributor to a small value tilt. Even if you have a strong tilt at 20% of total portfolio no single company in the small value index is anywhere near 1%.

The top two holdings in IJS as of 12/13/18 are Spire, Inc. at 1.12% followed by Wolverine at 0.85%. Divide by 5 if you overweight at 20% small value (plus a negligible extra for whats in Total Stock Market). Any single company in IJS makes a trivial difference.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 11:37 am

Random Walker wrote:
Sun Dec 16, 2018 11:20 am
I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave
Let's assume that idiosyncratic risk refers to single stock risk. However, IJS is a more concentrated investment than VTI, and concentration is not without risk.

The following is a contrived example, but illustrates that there is concentration risk. Assume every stock in IJS goes bankrupt, but no other stock in the US market does. A VTI owner has lost 1.91%; an IJS owner has lost everything.

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by DaufuskieNate » Sun Dec 16, 2018 11:44 am

Park wrote:
Sun Dec 16, 2018 11:11 am
VTI is concentrated in megacap/large cap stocks.

The following is from someone else, not me:

"As for the weaknesses of indexes: the megacap issue doesn't give me that much pause, because I view the megacaps as inherently more diversified; e.g. should you underweight Exxon vs a collection of small caps with the same number of oil fields in the aggregate?...Even something as concentrated as Apple has moving pieces, the iPhone markets often move very differently from one another"
Enron and more recently, JNJ, are evidence that large cap stocks are still subject to idiosyncratic risk.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 11:49 am

DaufuskieNate wrote:
Sun Dec 16, 2018 11:44 am
Park wrote:
Sun Dec 16, 2018 11:11 am
VTI is concentrated in megacap/large cap stocks.

The following is from someone else, not me:

"As for the weaknesses of indexes: the megacap issue doesn't give me that much pause, because I view the megacaps as inherently more diversified; e.g. should you underweight Exxon vs a collection of small caps with the same number of oil fields in the aggregate?...Even something as concentrated as Apple has moving pieces, the iPhone markets often move very differently from one another"
Enron and more recently, JNJ, are evidence that large cap stocks are still subject to idiosyncratic risk.
Any stock, large or small, has idiosyncratic risk. But with VTI, you've diversified away that idiosyncratic risk. With IJS, it's there; you can call it concentration risk instead, if you want.

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by DaufuskieNate » Sun Dec 16, 2018 12:02 pm

Park wrote:
Sun Dec 16, 2018 11:49 am
DaufuskieNate wrote:
Sun Dec 16, 2018 11:44 am
Park wrote:
Sun Dec 16, 2018 11:11 am
VTI is concentrated in megacap/large cap stocks.

The following is from someone else, not me:

"As for the weaknesses of indexes: the megacap issue doesn't give me that much pause, because I view the megacaps as inherently more diversified; e.g. should you underweight Exxon vs a collection of small caps with the same number of oil fields in the aggregate?...Even something as concentrated as Apple has moving pieces, the iPhone markets often move very differently from one another"
Enron and more recently, JNJ, are evidence that large cap stocks are still subject to idiosyncratic risk.
Any stock, large or small, has idiosyncratic risk. But with VTI, you've diversified away that idiosyncratic risk. With IJS, it's there; you can call it concentration risk instead, if you want.
This depends on your definition of concentration. To use a somewhat less contrived example, if JNJ goes bankrupt VTI loses 1.36%. (JNJ is the fifth largest holding.) If the fifth largest holding of IJS goes bankrupt, it loses .72%.

User avatar
Ben Mathew
Posts: 808
Joined: Tue Mar 13, 2018 11:41 am
Location: Seattle
Contact:

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Ben Mathew » Sun Dec 16, 2018 12:07 pm

Park wrote:
Sun Dec 16, 2018 11:37 am
The following is a contrived example, but illustrates that there is concentration risk. Assume every stock in IJS goes bankrupt, but no other stock in the US market does. A VTI owner has lost 1.91%; an IJS owner has lost everything.

The probability that all 470 companies in the IJS go bankrupt due to idiosyncratic risk is vanishingly small. It's possible that something like that could happen due to group level risk shared by all the small-value companies in IJS. But not due to idiosyncratic risk. Idiosyncratic risk is not really a concern once you get past, say, 50 or 100 stocks. The law of large numbers wipes it out.

I tilt towards small cap and value and I feel that I am taking on more risk than a market portfolio. But I'm worried only about factor level risk--maybe the shared characteristics that made small cap and value come out ahead in the past will perform badly in the future. An individual company's idiosyncratic risk does not worry me at all.

stan1
Posts: 8023
Joined: Mon Oct 08, 2007 4:35 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by stan1 » Sun Dec 16, 2018 12:23 pm

Again to OP: Are you looking at a domestic equity position that is 100% IJS (e.g. the "Larry Portfolio") or one that is let's say 70% VTI and 30% IJS. Very few people would choose to invest 100% of their domestic equity portfolio in SCV. Requires a lot of faith in theory to eschew Apple and the rest of the S&P 500 in their entirety. Have to look at the whole portfolio not just a part of it in isolation to do the thought analysis you are attempting.

User avatar
nisiprius
Advisory Board
Posts: 39761
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: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by nisiprius » Sun Dec 16, 2018 12:36 pm

I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.

But just because small value has higher risk doesn't mean anything, except to be aware of it and to make overall portfolio adjustments to compensate.

Now as to idiosyncratic risk, I thought that mean the special, unique risks of a single stock, the individual peculiarity of one single business encountering specific conditions. Something that might affect (say) Kraft, but not General Mills. I don't know why an index fund with 600 small-cap value stocks in it would have "idiosyncratic risk." If the only small-cap value companies in the world were all, say, shoestore chains, then maybe you'd have idiosyncratic risk, but that's not the case.
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.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 12:40 pm

nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.

But just because small value has higher risk doesn't mean anything, except to be aware of it and to make overall portfolio adjustments to compensate.

Now as to idiosyncratic risk, I thought that mean the risk of a single stock, the individual peculiarity of one single business encountering specific conditions that might affect (say) Kraft, but not General MIlls. I don't know why an index fund with 600 small-cap value stocks in it would have "idiosyncratic risk." If the only small-cap value companies in the world were all, say, shoestore chains, then maybe you'd have idiosyncratic risk, but that's not the case.
More than one poster has called into question my use of the expression idiosyncratic risk. With that in mind, feel free to substitute "concentration risk", wherever I've said idiosyncratic risk.

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by DaufuskieNate » Sun Dec 16, 2018 12:57 pm

I have a very strong tilt to small value in a globally diversified equity position. There are 7,809 individual stocks held by my funds. This is more than twice the number of stocks in VTI, and more than 15X the number of stocks in the S&P 500. This portfolio has risk, to be sure, but it's not concentration risk. For the record, I also do not consider the S&P 500 to have concentration risk. So, to answer the question posed by this thread, I believe it is entirely possible, even easy, to construct a portfolio with a small value factor tilt that does not increase idiosyncratic or concentration risk.

User avatar
nedsaid
Posts: 12976
Joined: Fri Nov 23, 2012 12:33 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by nedsaid » Sun Dec 16, 2018 1:15 pm

Random Walker wrote:
Sun Dec 16, 2018 11:20 am
I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave
I go with Dave on this one.
A fool and his money are good for business.

columbia
Posts: 2155
Joined: Tue Aug 27, 2013 5:30 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by columbia » Sun Dec 16, 2018 1:26 pm

nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.

But just because small value has higher risk doesn't mean anything, except to be aware of it and to make overall portfolio adjustments to compensate.

Now as to idiosyncratic risk, I thought that mean the special, unique risks of a single stock, the individual peculiarity of one single business encountering specific conditions. Something that might affect (say) Kraft, but not General Mills. I don't know why an index fund with 600 small-cap value stocks in it would have "idiosyncratic risk." If the only small-cap value companies in the world were all, say, shoestore chains, then maybe you'd have idiosyncratic risk, but that's not the case.

Most would agree that international is riskier for US investors. Seemingly, that should have rewarded investors with a premium over the long run. We know that hasn’t happened for 1970-2018.

I would be wary of anyone saying that a subset of equities will reward you with higher returns, provided that you are willing to take on more risk (vs TSM).

Markets don’t care about one’s calculated guesses.

lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 11:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by lack_ey » Sun Dec 16, 2018 1:29 pm

Factor tilts in most implementations should have little impact on idiosyncratic risk, especially through basically any passive or psuedo-passive DFA-type mutual fund or ETF. These run widely diversified, not concentrated allocations.

Concentrating in certain parts of the market will increase risk, but that is not idiosyncratic risk, Furthermore, certain factors line up with characteristics that are empirically riskier (size factor and smaller-cap stocks being riskier is an easy example), though others may not and may in fact have an opposite effect (low volatility, quality, etc.).

Overall I would be more specific about specifying "factor tilts" to evaluate the question more precisely. There's a difference between a portfolio of 10 stocks selected on value screens, some ~1000 stock fundamental-weighted ETF, an actively managed quality-focused mutual fund, a large cap multifactor fund targeting value/momentum/quality, a small cap multifactor fund targeting value/quality/low vol, etc.

Valuethinker
Posts: 39254
Joined: Fri May 11, 2007 11:07 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Valuethinker » Sun Dec 16, 2018 1:41 pm

Park wrote:
Sun Dec 16, 2018 12:40 pm
nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.

But just because small value has higher risk doesn't mean anything, except to be aware of it and to make overall portfolio adjustments to compensate.

Now as to idiosyncratic risk, I thought that mean the risk of a single stock, the individual peculiarity of one single business encountering specific conditions that might affect (say) Kraft, but not General MIlls. I don't know why an index fund with 600 small-cap value stocks in it would have "idiosyncratic risk." If the only small-cap value companies in the world were all, say, shoestore chains, then maybe you'd have idiosyncratic risk, but that's not the case.
More than one poster has called into question my use of the expression idiosyncratic risk. With that in mind, feel free to substitute "concentration risk", wherever I've said idiosyncratic risk.
I think your usage in Modern Portfolio Theory terms is clear?

Idiosyncratic risk is diversifiable risk i.e. if the fully diversified investor can diversify it away. Stock specific risk.

To address your question:

- yes you have higher idiosyncratic risk
-that probably shows up in 2 ways: 1). sectoral concentration 2). small stocks probably have less financial strength than large stocks (less access to capital markets)

In addition, there's no way owning a small bank in Georgia is the same as owning JP Morgan, even if they are in the same "sector". Even more so owning some small software company supplying niche business-to-business applications than owning Microsoft or Google.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 1:52 pm

nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.
The risk that you're alluding to is that factors, such as small and value, may not exist in the future. That risk doesn't bother me that much. If the factors disappear, then your return should be similar to the market return. If you've tilted in a cost and tax efficient manner, you shouldn't lose much compared to the market.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Sun Dec 16, 2018 2:02 pm

Valuethinker wrote:
Sun Dec 16, 2018 1:41 pm
Park wrote:
Sun Dec 16, 2018 12:40 pm
nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.

But just because small value has higher risk doesn't mean anything, except to be aware of it and to make overall portfolio adjustments to compensate.

Now as to idiosyncratic risk, I thought that mean the risk of a single stock, the individual peculiarity of one single business encountering specific conditions that might affect (say) Kraft, but not General MIlls. I don't know why an index fund with 600 small-cap value stocks in it would have "idiosyncratic risk." If the only small-cap value companies in the world were all, say, shoestore chains, then maybe you'd have idiosyncratic risk, but that's not the case.
More than one poster has called into question my use of the expression idiosyncratic risk. With that in mind, feel free to substitute "concentration risk", wherever I've said idiosyncratic risk.
I think your usage in Modern Portfolio Theory terms is clear?

Idiosyncratic risk is diversifiable risk i.e. if the fully diversified investor can diversify it away. Stock specific risk.

To address your question:

- yes you have higher idiosyncratic risk
-that probably shows up in 2 ways: 1). sectoral concentration 2). small stocks probably have less financial strength than large stocks (less access to capital markets)

In addition, there's no way owning a small bank in Georgia is the same as owning JP Morgan, even if they are in the same "sector". Even more so owning some small software company supplying niche business-to-business applications than owning Microsoft or Google.
I agree with you about sectoral concentration. That raised its head in 2008, when value portfolios had significant exposure to the financial sector. Sectoral concentration may be common feature of factor investing in general, whether value or momentum or quality etc. And I also agree with you about concentration risk in small caps manifesting as decreased financial strength.

HEDGEFUNDIE
Posts: 3973
Joined: Sun Oct 22, 2017 2:06 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by HEDGEFUNDIE » Sun Dec 16, 2018 5:08 pm

Park wrote:
Sun Dec 16, 2018 10:31 am
An argument made on this board is that factor tilts increase diversification. Factor tilted portfolios have exposure to the market factor (beta), but also exposure to other factors. Two popular factors here are small and value.

VTI (Vanguard Total Stock Market ETF) has 3641 stocks AVerage market cap is about $8.232 billion. Total market cap is about 29,972 billion.

As an example of a factor tilted fund, IJS (iShares S&) Small-Cap 600 Value ETF) has 462 stocks. Average market cap is about $1.24011 billion. Total market cap is about $573 billion.

IJS has 12.7% of the number of stocks that VTI has. If you assume that VTI represents 100% of the US investable stock market, then IJS is about 1.91% of the total US stock market by market cap.

IJS is much more concentrated than VTI. Aren't you taking on considerable idiosyncratic risk with factor tilted funds, such as IJS?
The largest holding in VTI is AAPL, which represents 3.4% of VTI. The top 10 holdings in VTI collectively represent 18.7% of VTI. 6 of the top 10 holdings are in Tech.

The largest holding in IJS is SR, which represents 1.1% of IJS. The top 10 holdings in IJS collectively represent 7.6% of IJS. The top sector in the top 10 holdings is Financials, which is only 4 of the top 10 holdings.

So you tell me, which one has more idiosyncratic/ concentration risk?

User avatar
JoMoney
Posts: 8176
Joined: Tue Jul 23, 2013 5:31 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by JoMoney » Sun Dec 16, 2018 5:58 pm

Idiosyncratic risk is risk that can be (theoretically) avoided through further diversification.
If the "risk" is that performance will deviate from the broad market average, then yes any tilt away from the market increases idiosyncratic risk.
A market weighted portfolio (however you want to define 'the market') holds a special position, in that it is Pareto optimal.
If the market is efficient, any deviations are simply moving along a spectrum of risk/return.
If the market is not efficient, deviating implies you have special information or an asymmetrical advantage (or that you're willing to risk being the patsy for those who do, in hopes of the chance at getting lucky).
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Sun Dec 16, 2018 6:18 pm

Park wrote:
Sun Dec 16, 2018 12:40 pm
nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

What tilters say is that if you look at the past statistics, despite having had higher risk, you could have used factor tilts to build portfolios that would have had the same return with less risk--or more return with the same risk--as total market portfolios. The controversial claim is that these statistical relationships are robust and persistent, and thus we can expect factor tilted portfolios to continue to have superior risk-adjusted return going forward.

But just because small value has higher risk doesn't mean anything, except to be aware of it and to make overall portfolio adjustments to compensate.

Now as to idiosyncratic risk, I thought that mean the risk of a single stock, the individual peculiarity of one single business encountering specific conditions that might affect (say) Kraft, but not General MIlls. I don't know why an index fund with 600 small-cap value stocks in it would have "idiosyncratic risk." If the only small-cap value companies in the world were all, say, shoestore chains, then maybe you'd have idiosyncratic risk, but that's not the case.
More than one poster has called into question my use of the expression idiosyncratic risk. With that in mind, feel free to substitute "concentration risk", wherever I've said idiosyncratic risk.
It’s been pointed out more than once that IJS is LESS concentrated than either the S&P 500 or total stock market, so I’m guessing “concentration” also isn’t the risk you’re trying to describe.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
Epsilon Delta
Posts: 8090
Joined: Thu Apr 28, 2011 7:00 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Epsilon Delta » Sun Dec 16, 2018 6:33 pm

JoMoney wrote:
Sun Dec 16, 2018 5:58 pm
A market weighted portfolio (however you want to define 'the market') holds a special position, in that it is Pareto optimal.
It's even more special than that. Not only is it Pareto optimal, but we know it's Pareto optimal.

stlutz
Posts: 5508
Joined: Fri Jan 02, 2009 1:08 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by stlutz » Sun Dec 16, 2018 6:50 pm

nisiprius wrote:
Sun Dec 16, 2018 12:36 pm
I'm not sure what you mean by idiosyncratic risk.

In the past, small-caps have consistently had more risk, by various measures, than the market as a whole; and value stocks have had more risk, by various measures than the market as a whole. I don't think that's seriously in dispute.

I think the thing to emphasize here is "market as a whole".

What is in dispute (at least with me) is whether small caps (say, the bottom 10% of the market) have more risk than any other slice of 10% of the market--could be the largest 4 companies overall; could be companies that start with the letter A. Personally, I would find a small cap fund to be less risky than just owning the largest 4 companies.

I would also dispute that the value half of the market is riskier than than the growth half of the market. If anything, statistically the opposite has proven to be somewhat the case, at least over the past 4 decades.

Bottom line is, owning 100% of the market will be less risky than owning any small segment you might pick out. In that sense, picking smaller slices does expose you to more random type of risks than just whether the economy as a whole is profitable or not.

typical.investor
Posts: 1363
Joined: Mon Jun 11, 2018 3:17 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by typical.investor » Sun Dec 16, 2018 6:51 pm

JoMoney wrote:
Sun Dec 16, 2018 5:58 pm
Idiosyncratic risk is risk that can be (theoretically) avoided through further diversification.
If the "risk" is that performance will deviate from the broad market average, then yes any tilt away from the market increases idiosyncratic risk.
A market weighted portfolio (however you want to define 'the market') holds a special position, in that it is Pareto optimal.
If the market is efficient, any deviations are simply moving along a spectrum of risk/return.
If the market is not efficient, deviating implies you have special information or an asymmetrical advantage (or that you're willing to risk being the patsy for those who do, in hopes of the chance at getting lucky).
I agree. Idiosyncratic risk might not be exactly the right term, nor is concentration risk, but I do believe concentrating your portfolio in a subset of stocks with a particular characteristic is exposing you to the risk that the characteristic doesn't do well in your holding period.

I wouldn't exactly call it concentration risk because you are well diversified over a number of stocks in different sectors (although you do have a sector tilt - underweight technology/healthcare and overweight financials [relative to other small cap stocks]).

Small value stocks typically do better in bursts, don't they? If you concentrate your holdings in them and hold during one of those periods - bingo. No guarantee that will happen in your holding period though.

But no, I don't think small value is idiosyncratic risk. It's not like you are holding a few companies and if something peculiar happens (CEO goes nuts, product has fluke default, etc etc).

fennewaldaj
Posts: 815
Joined: Sun Oct 22, 2017 11:30 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by fennewaldaj » Sun Dec 16, 2018 10:27 pm

Park wrote:
Sun Dec 16, 2018 1:52 pm


The risk that you're alluding to is that factors, such as small and value, may not exist in the future. That risk doesn't bother me that much. If the factors disappear, then your return should be similar to the market return. If you've tilted in a cost and tax efficient manner, you shouldn't lose much compared to the market.
This is not necessarily true though right? It is at least conceivable it could be come a negative factor ( a performance drag compared to the rest of the market) in the future. This could happen if small value got bid up relative to what it "should cost" due to the known anomaly. I know small values cost compared to the rest of the market is not currently more expensive than its history so it seems unlikely to me. The way that it could happen is if those people claiming that the large caps are becoming more and more dominant/monoplistic ect are right and small value should be cheaper than it is. Basically the "its different this time narrative" might be true. Now I don't think this is the case but is is certainly possible.

User avatar
packer16
Posts: 1306
Joined: Sat Jan 04, 2014 2:28 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by packer16 » Sun Dec 16, 2018 10:34 pm

I am not sure if we know that SCV is cheaper than history as we do not know the relative growth of the historical SCV firms versus the current SCV firms. If the market is getting more efficient then the relative growth rate should decline. I have not seen a relative growth analysis so IMO we do not know if we are more expensive than the past with apples to apples growth rates.

Packer
Buy cheap and something good might happen

fennewaldaj
Posts: 815
Joined: Sun Oct 22, 2017 11:30 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by fennewaldaj » Sun Dec 16, 2018 10:38 pm

vineviz wrote:
Sun Dec 16, 2018 6:18 pm


It’s been pointed out more than once that IJS is LESS concentrated than either the S&P 500 or total stock market, so I’m guessing “concentration” also isn’t the risk you’re trying to describe.
Each of those companies likely faces more idiosyncratic risk though right? Of course both funds have enough stocks to avoid much in the way of idiosyncratic risk. And I am honestly not sure if a portfolio with say 40% IJS 60% VTI has more of less idiosyncratic risk than 100% VTI. I know I feel more comfortable having less concentration in mega caps but that may just be a physiological thing for me.

fennewaldaj
Posts: 815
Joined: Sun Oct 22, 2017 11:30 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by fennewaldaj » Sun Dec 16, 2018 10:42 pm

packer16 wrote:
Sun Dec 16, 2018 10:34 pm
I am not sure if we know that SCV is cheaper than history as we do not know the relative growth of the historical SCV firms versus the current SCV firms. If the market is getting more efficient then the relative growth rate should decline. I have not seen a relative growth analysis so IMO we do not know if we are more expensive than the past with apples to apples growth rates.

Packer
I was thinking of several Larry Swedroe articles mention by Cliff Assness in a podcast and the JP Morgan guild to the market. I have not looked at the data my self. I haven't read up on EPS growth in SCV compared to the past.

https://am.jpmorgan.com/us/en/asset-man ... ets/viewer

User avatar
packer16
Posts: 1306
Joined: Sat Jan 04, 2014 2:28 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by packer16 » Sun Dec 16, 2018 11:04 pm

fennewaldaj wrote:
Sun Dec 16, 2018 10:42 pm
packer16 wrote:
Sun Dec 16, 2018 10:34 pm
I am not sure if we know that SCV is cheaper than history as we do not know the relative growth of the historical SCV firms versus the current SCV firms. If the market is getting more efficient then the relative growth rate should decline. I have not seen a relative growth analysis so IMO we do not know if we are more expensive than the past with apples to apples growth rates.

Packer
I was thinking of several Larry Swedroe articles mention by Cliff Assness in a podcast and the JP Morgan guild to the market. I have not looked at the data my self. I haven't read up on EPS growth in SCV compared to the past.

https://am.jpmorgan.com/us/en/asset-man ... ets/viewer
You can do the expensive/cheap analysis if you have the same set of companies in the sample but for SCV you have an ever changing group of firms who have different growth rates over time. If the market is getting more efficient over time the SCV sample should have lower growth rates over time. Part of the reason SCV has done well in the past is the actual growth rates achieved were higher than those implied in their prices. Now if folks are buying SCV stocks that are priced correctly then you are OK but if the demand for SCV has overpriced the stock relative to growth then you will underperform.

Packer
Buy cheap and something good might happen

fennewaldaj
Posts: 815
Joined: Sun Oct 22, 2017 11:30 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by fennewaldaj » Sun Dec 16, 2018 11:18 pm

packer16 wrote:
Sun Dec 16, 2018 11:04 pm
fennewaldaj wrote:
Sun Dec 16, 2018 10:42 pm
packer16 wrote:
Sun Dec 16, 2018 10:34 pm
I am not sure if we know that SCV is cheaper than history as we do not know the relative growth of the historical SCV firms versus the current SCV firms. If the market is getting more efficient then the relative growth rate should decline. I have not seen a relative growth analysis so IMO we do not know if we are more expensive than the past with apples to apples growth rates.

Packer
I was thinking of several Larry Swedroe articles mention by Cliff Assness in a podcast and the JP Morgan guild to the market. I have not looked at the data my self. I haven't read up on EPS growth in SCV compared to the past.

https://am.jpmorgan.com/us/en/asset-man ... ets/viewer
You can do the expensive/cheap analysis if you have the same set of companies in the sample but for SCV you have an ever changing group of firms who have different growth rates over time. If the market is getting more efficient over time the SCV sample should have lower growth rates over time. Part of the reason SCV has done well in the past is the actual growth rates achieved were higher than those implied in their prices. Now if folks are buying SCV stocks that are priced correctly then you are OK but if the demand for SCV has overpriced the stock relative to growth then you will underperform.

Packer
This is a long article and actually focused on large value but worth a read if the topic interest you. Basically they break down the S+P 500 by quintiles based on P/E and then look at eps. Basically they are cheap to start and then eps revert faster than they "should"

https://osam.com/Commentary/factors-from-scratch

Then in this breaks down eps growth in since 2010. It has not been pretty for value in the large cap space

https://osam.com/Commentary/osam-quarte ... er-q2-2018

So this is at least suggestive that eps has been bad for value companies hence poor returns. I don't know of any breakdown in the SC space though.

Northern Flicker
Posts: 5070
Joined: Fri Apr 10, 2015 12:29 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Northern Flicker » Mon Dec 17, 2018 12:47 am

Random Walker wrote:
Sun Dec 16, 2018 11:20 am
I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave
Sector risk is also a diversifiable risk. We can play a semantic game whether or not it belongs as part of idiosyncratic risk, but either way it should be diversified away.

Nobody knows what the optimal diversification might be, but the total US market has become more concentrated in the tech and financial sectors in the last 10-20 years. A small-cap value tilt may increase exposure to the financial sector but decrease exposure to the tech sector at present. It is theoretically possible the the mix that balances the overall size of the two sectors is optimal but that is no more established than any other mix being optimal.
Index fund investor since 1987.

columbia
Posts: 2155
Joined: Tue Aug 27, 2013 5:30 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by columbia » Mon Dec 17, 2018 12:51 am

jalbert wrote:
Mon Dec 17, 2018 12:47 am
Random Walker wrote:
Sun Dec 16, 2018 11:20 am
I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave
Sector risk is also a diversifiable risk. We can play a semantic game whether or not it belongs as part of idiosyncratic risk, but either way it should be diversified away.

Nobody knows what the optimal diversification might be, but the total US market has become more concentrated in the tech and financial sectors in the last 10-20 years. A small-cap value tilt may increase exposure to the financial sector but decrease exposure to the tech sector at present. It is theoretically possible the the mix that balances the overall size of the two sectors is optimal but that is no more established than any other mix being optimal.
ProShrares is happy to sell you an ex-tech ETF (without having slum in the SC space):
http://www.proshares.com/funds/spxt.html

typical.investor
Posts: 1363
Joined: Mon Jun 11, 2018 3:17 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by typical.investor » Mon Dec 17, 2018 2:40 am

Park wrote:
Sun Dec 16, 2018 10:31 am
An argument made on this board is that factor tilts increase diversification. Factor tilted portfolios have exposure to the market factor (beta), but also exposure to other factors. Two popular factors here are small and value.
If you take the market and excluded some stocks (say large growth), then the resulting portfolio has loadings on beta, value and size.

What loadings do the excluded stocks have? It can't be only beta.

Obviously, large growth stocks will have a negative loading on size and value.

If size and value returns are negative in your holding period, a small value tilt will suffer the consequences of having a concentrated portfolio. The market will outperform it and those portfolios holding a large growth concentration will outperform the market.

It seems obvious that the market is better diversified than either a small value tilted or a large growth tilted portfolio, and this is due to it's not having a net exposure to anything other than the market.

Factor proponents simply assume that your factor returns from a tilted portfolio are destined to outperform in your holding period, and so claim better diversification on the (assumed) "drivers of return". It's a false claim though and actually factor returns could be negative.

So yes, factor tilted portfolios are more diversified provided we make assumptions which aren't necessarily true. Diversification for market portfolios, on the other hand, don't depend on any particular factors outperforming in your holding period.
Last edited by typical.investor on Mon Dec 17, 2018 4:30 am, edited 1 time in total.

User avatar
asset_chaos
Posts: 1701
Joined: Tue Feb 27, 2007 6:13 pm
Location: Melbourne

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by asset_chaos » Mon Dec 17, 2018 4:06 am

Providers of factor portfolios have to choose what index or rules to follow to get the factor exposure, including roughly how many holdings. They arrive at markedly different answers:

Small-Cap 600 Value has 462 holdings, and 15 year Sharpe ratio 0.52

Vanguard small cap value in index fund has 870, 0.52

DFA US targeted value has 1513, 0.49

IJS quoted from OP and other two from M* as of a few moments ago. A factor of 3 between IJS and the DFA fund seems like a notable difference. It must be extra effort for DFA to run 3x the stocks as IJS; presumably they discern some benefit. I think it would matter if most of the extra risk or extra return came persistently from a small subset of the factor universe of companies, as it does for the market factor. In which case more companies would be better in order to be more likely to generate the premium. I don't know if that's the case for factors in general.

15 year Sharpe ratios from M* don't support, though, concluding there is much difference in risk adjusted return between these three funds. The slightly lower Sharpe ratio for the DFA fund was due to it having been a little more volatile for essentially the same return as the other two over the time span.

(AS someone will ask, M* lists the 15 year Sharpe ratio for total stock index fund vtsmx as 0.60 with about a quarter percent a year less return than these small value funds but with much less volatility.)
Regards, | | Guy

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Mon Dec 17, 2018 6:58 am

jalbert wrote:
Mon Dec 17, 2018 12:47 am
Random Walker wrote:
Sun Dec 16, 2018 11:20 am
I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave
Sector risk is also a diversifiable risk. We can play a semantic game whether or not it belongs as part of idiosyncratic risk, but either way it should be diversified away.

Nobody knows what the optimal diversification might be, but the total US market has become more concentrated in the tech and financial sectors in the last 10-20 years. A small-cap value tilt may increase exposure to the financial sector but decrease exposure to the tech sector at present. It is theoretically possible the the mix that balances the overall size of the two sectors is optimal but that is no more established than any other mix being optimal.
Historically, the optimal allocation to US small cap stocks from a diversification perspective has been pretty stable about 43% within a reasonably narrow range of +/- 5%.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Mon Dec 17, 2018 7:07 am

typical.investor wrote:
Mon Dec 17, 2018 2:40 am
It seems obvious that the market is better diversified than either a small value tilted or a large growth tilted portfolio, and this is due to it's not having a net exposure to anything other than the market.
This might seem obvious, but it's not true. As has been demonstrated, both empirically and theoretically, countless times in discussions before.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

typical.investor
Posts: 1363
Joined: Mon Jun 11, 2018 3:17 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by typical.investor » Mon Dec 17, 2018 8:47 am

vineviz wrote:
Mon Dec 17, 2018 7:07 am
typical.investor wrote:
Mon Dec 17, 2018 2:40 am
It seems obvious that the market is better diversified than either a small value tilted or a large growth tilted portfolio, and this is due to it's not having a net exposure to anything other than the market.
This might seem obvious, but it's not true. As has been demonstrated, both empirically and theoretically, countless times in discussions before.
Well, yes, it has been shown that if ex-ante factor returns are assumed to be the same as what we see ex-post, then a factor tilted portfolio will be more diversified over the drivers of return.

If, however, your investment period is like the last decade, you actually are less diversified over the drivers of return.

I don't assume ex-ante equates ex-post, so can't share the view that factor portfolios are more diversified.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Mon Dec 17, 2018 8:59 am

typical.investor wrote:
Mon Dec 17, 2018 8:47 am
vineviz wrote:
Mon Dec 17, 2018 7:07 am
typical.investor wrote:
Mon Dec 17, 2018 2:40 am
It seems obvious that the market is better diversified than either a small value tilted or a large growth tilted portfolio, and this is due to it's not having a net exposure to anything other than the market.
This might seem obvious, but it's not true. As has been demonstrated, both empirically and theoretically, countless times in discussions before.
Well, yes, it has been shown that if ex-ante factor returns are assumed to be the same as what we see ex-post, then a factor tilted portfolio will be more diversified over the drivers of return.

If, however, your investment period is like the last decade, you actually are less diversified over the drivers of return.

I don't assume ex-ante equates ex-post, so can't share the view that factor portfolios are more diversified.
You're suggesting that we should only own the assets with the highest future returns, and that diversification should be avoided if it lowers returns in any way?
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

typical.investor
Posts: 1363
Joined: Mon Jun 11, 2018 3:17 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by typical.investor » Mon Dec 17, 2018 9:13 am

vineviz wrote:
Mon Dec 17, 2018 8:59 am
typical.investor wrote:
Mon Dec 17, 2018 8:47 am
vineviz wrote:
Mon Dec 17, 2018 7:07 am
typical.investor wrote:
Mon Dec 17, 2018 2:40 am
It seems obvious that the market is better diversified than either a small value tilted or a large growth tilted portfolio, and this is due to it's not having a net exposure to anything other than the market.
This might seem obvious, but it's not true. As has been demonstrated, both empirically and theoretically, countless times in discussions before.
Well, yes, it has been shown that if ex-ante factor returns are assumed to be the same as what we see ex-post, then a factor tilted portfolio will be more diversified over the drivers of return.

If, however, your investment period is like the last decade, you actually are less diversified over the drivers of return.

I don't assume ex-ante equates ex-post, so can't share the view that factor portfolios are more diversified.
You're suggesting that we should only own the assets with the highest future returns, and that diversification should be avoided if it lowers returns in any way?
Quite the mischaracterization on your part, isn't it!

I have said nothing of the sort whatsoever.

A more realistic view of what I said is that while some factors (such as value, size, mom, qual, and low beta) may have higher expected returns, a market portfolio is more diversified because it also includes holdings which don't load on those factors. Because the factors with highest expected returns (and usually most risk) can experience long periods of underperformance, have shifting correlations that can converge, have quite non -normal return distributions and have (factor) momentum which can prolong down times, that a factor tilted portfolio is less diversified because it will do better or worse than the market depending on the period.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Mon Dec 17, 2018 9:53 am

typical.investor wrote:
Mon Dec 17, 2018 9:13 am
Quite the mischaracterization on your part, isn't it!

I have said nothing of the sort whatsoever.

A more realistic view of what I said is that while some factors (such as value, size, mom, qual, and low beta) may have higher expected returns, a market portfolio is more diversified because it also includes holdings which don't load on those factors. Because the factors with highest expected returns (and usually most risk) can experience long periods of underperformance, have shifting correlations that can converge, have quite non -normal return distributions and have (factor) momentum which can prolong down times, that a factor tilted portfolio is less diversified because it will do better or worse than the market depending on the period.
It's really hard to characterize what you've said because there are so many conflicting statements and non-standard usage of financial terms. But the last part of your last sentence says that the measure of whether portfolio is diversified is whether or not it does "better or worse than the market".

This is a very idiosyncratic view of the role of diversification.

For one thing, the goal of diversification is NOT to beat the market but rather is to manage the risk of the portfolio.

For a second thing, this formulation is based on an implicit assumption that the market portfolio is the maximally diversified portfolio. Not only is this demonstrably NOT true, it's logically flawed: you're assuming the very thing that you are attempting to prove in order to prove it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Mon Dec 17, 2018 10:23 am

vineviz wrote:
Mon Dec 17, 2018 6:58 am
jalbert wrote:
Mon Dec 17, 2018 12:47 am
Random Walker wrote:
Sun Dec 16, 2018 11:20 am
I do not think factor tilts increase idiosyncratic risk. Idiosyncratic risk generally is single stock risk. Most all the factor tilting discussed here is with cap weighted, passive/index funds that have hundreds or thousands of individual stocks. Single stock risk is effectively eliminated.

Dave
Sector risk is also a diversifiable risk. We can play a semantic game whether or not it belongs as part of idiosyncratic risk, but either way it should be diversified away.

Nobody knows what the optimal diversification might be, but the total US market has become more concentrated in the tech and financial sectors in the last 10-20 years. A small-cap value tilt may increase exposure to the financial sector but decrease exposure to the tech sector at present. It is theoretically possible the the mix that balances the overall size of the two sectors is optimal but that is no more established than any other mix being optimal.
Historically, the optimal allocation to US small cap stocks from a diversification perspective has been pretty stable about 43% within a reasonably narrow range of +/- 5%.
Could I ask the source of your data for the last remark?

randomguy
Posts: 8531
Joined: Wed Sep 17, 2014 9:00 am

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by randomguy » Mon Dec 17, 2018 10:24 am

Ben Mathew wrote:
Sun Dec 16, 2018 12:07 pm
Park wrote:
Sun Dec 16, 2018 11:37 am
The following is a contrived example, but illustrates that there is concentration risk. Assume every stock in IJS goes bankrupt, but no other stock in the US market does. A VTI owner has lost 1.91%; an IJS owner has lost everything.

The probability that all 470 companies in the IJS go bankrupt due to idiosyncratic risk is vanishingly small. It's possible that something like that could happen due to group level risk shared by all the small-value companies in IJS. But not due to idiosyncratic risk. Idiosyncratic risk is not really a concern once you get past, say, 50 or 100 stocks. The law of large numbers wipes it out.

I tilt towards small cap and value and I feel that I am taking on more risk than a market portfolio. But I'm worried only about factor level risk--maybe the shared characteristics that made small cap and value come out ahead in the past will perform badly in the future. An individual company's idiosyncratic risk does not worry me at all.
What about if we reverse the question. the 470 largest companies go bankrupt. VTI loses like 80% of its value while IJS loses 0%. Is that really any different? Yes losing 85% is better than 100% but that is a pretty minimal difference to me. Obviously both are pretty far out there possibilities:)

I would worry a lot more about the risk of underperformance. I haven't looked recently but IJS tends to be underweight in Technology and healthcare while holding more financial and industrial stocks. Who the heck knows if that is a good bet going forward.

User avatar
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Mon Dec 17, 2018 11:05 am

Park wrote:
Mon Dec 17, 2018 10:23 am
Historically, the optimal allocation to US small cap stocks from a diversification perspective has been pretty stable about 43% within a reasonably narrow range of +/- 5%.
Could I ask the source of your data for the last remark?
I used Portfolio Visualizer's Rolling Portfolio Optimization tool compute the "Maximize Diversification" portfolio each month on a rolling 36-month loopback for VFINX (Vanguard 500 Index fund) and NAESX (Vanguard Small Cap Index), but you get nearly identical results using other proxy funds and methods. Using a small cap value fund instead of NAESX would very slightly decrease the allocation to VFINX, but it's well within the confidence interval.

https://www.portfoliovisualizer.com/rol ... bol2=NAESX

Note that this tool will show a 100% cash portfolio for the "optimized" portfolio if the trailing return is negative so you have to either manually calculate the portfolio for those periods or use the compared allocation (i.e. "Inverse Volatility").
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
nisiprius
Advisory Board
Posts: 39761
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: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by nisiprius » Mon Dec 17, 2018 8:37 pm

Park wrote:
Mon Dec 17, 2018 10:23 am
vineviz wrote:
Mon Dec 17, 2018 6:58 am
Historically, the optimal allocation to US small cap stocks from a diversification perspective has been pretty stable about 43% within a reasonably narrow range of +/- 5%.
Could I ask the source of your data for the last remark?
It is more important to ask for the definition of "diversification." Vineviz has his own favorite definition of "diversification."

There has been more than one sterile discussion following the pattern:
"Your portfolio is less diversified than mine."
"No, it isn't, according to my definition of 'diversified.'"
"Yes, it is, according to my definition of 'diversified.'"

For the record, and without using the term "diversified," and without any claims that MPT is good for anything... over the time period 1926-2017, the optimum portfolio, represented by the tangent to the efficient frontier, of SBBI Large-Company Stocks (S&P 500 and predecessors) and SBBI Small-Company Stocks (Banz 1980 data and successors) was 74% large-caps, 26% small-caps... and the improvement over 100% large-caps, shown graphically by the difference between the red and yellow lines, was tiny.

Image

Over thirty-year periods of time, there have been periods (1956-1985) when the optimum was 100% small-caps,

Image

and periods (1926-1955) when it was 100% large-caps.

Image

Whether there has been a "optimal allocation" that has been "pretty stable" depends on what lens you are looking through.
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
vineviz
Posts: 5391
Joined: Tue May 15, 2018 1:55 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by vineviz » Mon Dec 17, 2018 10:12 pm

nisiprius wrote:
Mon Dec 17, 2018 8:37 pm
For the record, and without using the term "diversified," and without any claims that MPT is good for anything...
I'm glad you refrained from using the term "diversified" because nothing in your post had anything to do with diversification.

Which makes me wonder why you posted it reply to my comment which WAS discussing the topic of diversification.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Mon Dec 17, 2018 11:33 pm

packer16 wrote:
Sun Dec 16, 2018 11:04 pm
fennewaldaj wrote:
Sun Dec 16, 2018 10:42 pm
packer16 wrote:
Sun Dec 16, 2018 10:34 pm
I am not sure if we know that SCV is cheaper than history as we do not know the relative growth of the historical SCV firms versus the current SCV firms. If the market is getting more efficient then the relative growth rate should decline. I have not seen a relative growth analysis so IMO we do not know if we are more expensive than the past with apples to apples growth rates.

Packer
I was thinking of several Larry Swedroe articles mention by Cliff Assness in a podcast and the JP Morgan guild to the market. I have not looked at the data my self. I haven't read up on EPS growth in SCV compared to the past.

https://am.jpmorgan.com/us/en/asset-man ... ets/viewer
You can do the expensive/cheap analysis if you have the same set of companies in the sample but for SCV you have an ever changing group of firms who have different growth rates over time. If the market is getting more efficient over time the SCV sample should have lower growth rates over time. Part of the reason SCV has done well in the past is the actual growth rates achieved were higher than those implied in their prices. Now if folks are buying SCV stocks that are priced correctly then you are OK but if the demand for SCV has overpriced the stock relative to growth then you will underperform.

Packer
I'm going to talk about small value tilts, but the following probably applies to factor tilts in general. I hadn't thought about the possibility of SCV becoming overpriced, but it should be considered. IJS is about 2% of the US stock market. As shown in 1999, the market in general can get grossly overpriced. If that can happen to the market, it can certainly happen to 2% of the market.

There are literally hundreds (thousands?) of hedge fund that are very aware of small value tilts. DFA has about $600 billion in assets, and from what I can see, tilts to small and value play an important role in how those assets are invested. At the level of the retail investor, this board shows that such tilts are not unpopular. My perception is that the interest in small growth investing is not equal to that of small value investing.

One argument I hear against overpricing is that small cap value should have done very well in the last 15 years, if it was becoming overpriced. Also, if small cap value was becoming overpriced, the disparity between in price ratios between growth and value stocks should have lessened. But those two arguments assume that the market is not becoming more efficient. If the market is pricing SCV stocks better than it did 15 years ago, I wouldn't expect small cap value to have done well in the last 15 years or that the disparity in price ratios between growth and value stocks should have lessened.

If the value premium is risk based, then it shouldn't decline with time. Frankly, I've never heard compelling arguments for the value premium being risk based. I have heard such arguments for the small premium. Ironically, it looks like the small premium hasn't done well since about 1985, when its existence became widely known.

If the value premium is behavioral in its origin, then it could decline with time. People talk about limits to arbitrage preventing behavioral mistakes from being exploited. But the popularity of SCV could create a behavioral mistake that negates the behavioral mistakes of the previous sentence.

I've been on this board long enough to hear people advocate REITs or commodity investing; those voices have been quieter in the last few years. Will SCV go down the same road?

If I had to hazard a guess, I think SCV premia will persist, although less than in the past. Bubbles have always existed, and unless human behavior changes, always will. In bubbles, growth becomes overpriced. Bubbles occur infrequently, but the behavior that creates them is likely always there. It's just that in bubbles, that behavior assumes greater importance than usual. And that behavior will likely manifest itself chronically as value outperforming growth.

fennewaldaj
Posts: 815
Joined: Sun Oct 22, 2017 11:30 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by fennewaldaj » Mon Dec 17, 2018 11:47 pm

Park wrote:
Mon Dec 17, 2018 11:33 pm


One argument I hear against overpricing is that small cap value should have done very well in the last 15 years, if it was becoming overpriced. Also, if small cap value was becoming overpriced, the disparity between in price ratios between growth and value stocks should have lessened. But those two arguments assume that the market is not becoming more efficient. If the market is pricing SCV stocks better than it did 15 years ago, I wouldn't expect small cap value to have done well in the last 15 years or that the disparity in price ratios between growth and value stocks should have lessened.
All other things equal if SCV is priced more efficiently then the value growth spreads should be lessened right? In the past SCV was systematically underpriced. Now as I noted above it is possible that secular forces are actually making SCV companies worth less (Large capst are taking over the world arguments). If that was the case it would be possible for relative valuations to stay the same while SCV is getting overpriced.

Topic Author
Park
Posts: 691
Joined: Sat Nov 06, 2010 4:56 pm

Re: Do Factor Tilts, Such As Small Value, Increase Idiosyncratic Risk?

Post by Park » Tue Dec 18, 2018 8:52 am

fennewaldaj wrote:
Mon Dec 17, 2018 11:47 pm
Park wrote:
Mon Dec 17, 2018 11:33 pm


One argument I hear against overpricing is that small cap value should have done very well in the last 15 years, if it was becoming overpriced. Also, if small cap value was becoming overpriced, the disparity between in price ratios between growth and value stocks should have lessened. But those two arguments assume that the market is not becoming more efficient. If the market is pricing SCV stocks better than it did 15 years ago, I wouldn't expect small cap value to have done well in the last 15 years or that the disparity in price ratios between growth and value stocks should have lessened.
All other things equal if SCV is priced more efficiently then the value growth spreads should be lessened right? In the past SCV was systematically underpriced. Now as I noted above it is possible that secular forces are actually making SCV companies worth less (Large capst are taking over the world arguments). If that was the case it would be possible for relative valuations to stay the same while SCV is getting overpriced.
There are those on this board, with much greater knowledge than mine, so I may very well get corrected on the following.

The good majority of small cap value stocks deserve their valuations. It's a comparatively small group that have resulted in the increased return. If you could select that small group, returns of SCV would decrease, but valuations wouldn't change that much. About being able to select that small group, there are straightforward ways to do that, such as the Piotroski score. For those good at fundamental analysis, there would be other ways.

About SCV and risk, IJS has about 13% of the stocks that VTI has. When you have 87% less stocks, you are less diversified. When you are less diversified, you are taking on an increased risk. That risk may not be called idiosyncratic or concentration risk, and SCV may have features that outweigh that risk. But there is an increased risk associated with having 87% less stocks.
Last edited by Park on Tue Dec 18, 2018 8:59 am, edited 1 time in total.

Post Reply