how much tilted is 'a little' or 'a lot'?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
letsgobobby
Posts: 11482
Joined: Fri Sep 18, 2009 1:10 am

how much tilted is 'a little' or 'a lot'?

Post by letsgobobby » Fri Feb 15, 2013 8:38 pm

How much tilting to small or value must one have to be meaningful? I see people tilt '10%' all the way up to a Swedroe or barbell type portfolio where 100% of stocks are small, value, or both.

In my case 1/3 of my stocks are small, value, or both, for both domestic and international. Is that a lot? A little? Is there a consensus?

livesoft
Posts: 62294
Joined: Thu Mar 01, 2007 8:00 pm

Re: how much tilted is 'a little' or 'a lot'?

Post by livesoft » Fri Feb 15, 2013 8:42 pm

Show us examples portfolios of what you mean by 10% tilt or 100% tilt please. I consider 0% tilt to be Total Stock Market Index.

Would you say a portfolio of 100% TSM is 0% tilted?
How about a portfolio of 100% Small-cap-value? Is that 100% tilted?

As for meaningful (whatever that means), my rule of thumb is to have between 45% and 50% of equities in the large-cap row of a Morningstar X-ray 9-box style grid. If one has more large-caps than that, then one should tilt more to small/mid-caps.
Wiki This signature message sponsored by sscritic: Learn to fish.

EDN
Posts: 528
Joined: Sat Dec 29, 2012 7:55 pm
Location: Oklahoma City, OK

Tilted Portfolio Examples

Post by EDN » Fri Feb 15, 2013 9:34 pm

letsgobobby wrote:How much tilting to small or value must one have to be meaningful? I see people tilt '10%' all the way up to a Swedroe or barbell type portfolio where 100% of stocks are small, value, or both.

In my case 1/3 of my stocks are small, value, or both, for both domestic and international. Is that a lot? A little? Is there a consensus?
There is no consensus on this, but we can speak in generalities. The DFA US Small Value fund has a 0.8 exposure to the size premium and about 0.7 to the value premium. On a broadly diversified basis, it's difficult to get more targeted exposure to the return dimensions than that. So something in the 0.6 to 0.8 should be labeled as "extreme". A portfolio with 0.3 to 0.4 is probably considered "moderate", and 0.2 or less is very mild.

You probably want to have at least a 0.2 or so tilt to small and value or it won't be enough to derive any benefit.

What's this in allocation terms? Extreme = 100% S&P 600 Value. Moderate = 40% TSM, 60% S&P 600 Value. Mild = 70% TSM, 30% S&P 600 Value.

Eric

stlutz
Posts: 4660
Joined: Fri Jan 02, 2009 1:08 am

Re: how much tilted is 'a little' or 'a lot'?

Post by stlutz » Fri Feb 15, 2013 9:41 pm

I for one would not consider allocating 60% of your equity position on approx 2% of the market to be a "moderate" tilt. That's a pretty serious commitment to an expectation that one's forecast of future stock returns will work out.

User avatar
nisiprius
Advisory Board
Posts: 36467
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: how much tilted is 'a little' or 'a lot'?

Post by nisiprius » Fri Feb 15, 2013 9:46 pm

In theory you are supposed calculate an optimum percentage, based on the standard deviations of the asset classes and the correlations with each other. There's an optimum point on the "efficient frontier." If the standard deviations and correlations don't change, the optimum percentages doesn't change.

But's not supposed to be a matter of gut instinct or taste, it's supposed to be a calculated optimum. Since I don't tilt, I've never looked much into what that calculated optimum is supposed to be. I assume that Bill Schultheis' Coffeehouse Portfolio isn't far off the mark even though he uses round-number percentages.

Of course correlations and standard deviations do change--correlations in particular are wildly unstable. So the mix that was the optimum for the past isn't necessarily the optimum going forward.
Last edited by nisiprius on Fri Feb 15, 2013 9:48 pm, edited 1 time in total.
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
ruralavalon
Posts: 13728
Joined: Sat Feb 02, 2008 10:29 am
Location: Illinois

Re: how much tilted is 'a little' or 'a lot'?

Post by ruralavalon » Fri Feb 15, 2013 9:48 pm

letsgobobby wrote: In my case 1/3 of my stocks are small, value, or both, for both domestic and international. Is that a lot? A little? Is there a consensus?
Don't think that there is really a "consensus". I would consider something like a 70:30 ratio of TSM to SCV as a good choice, as suggested in R. Ferri's All About Asset Allocation. If this is "mild" thats OK by me; it is what he calcuates as optimum. As noted correlations change over time, and his suggestion is based on past performance.

Personally, we use less SCV than that.
"Everything should be as simple as it is, but not simpler." - Albert Einstein | Wiki article link:Getting Started

EDN
Posts: 528
Joined: Sat Dec 29, 2012 7:55 pm
Location: Oklahoma City, OK

Misconceptions of S&P 600 Value

Post by EDN » Fri Feb 15, 2013 10:04 pm

stlutz wrote:I for one would not consider allocating 60% of your equity position on approx 2% of the market to be a "moderate" tilt. That's a pretty serious commitment to an expectation that one's forecast of future stock returns will work out.
The S&P 600 Value, which is not particularly small or valuey, doesn't represent 2% of the market, you've miscalculated that. And that's just the US part of an equity portfolio anyway, not the entire allocation. Anyway, because TSM is dominated by large and mega cap stocks, you need a lot of a small value index that isn't that small or valuey to get healthy size and value exposure.

Eric

EDN
Posts: 528
Joined: Sat Dec 29, 2012 7:55 pm
Location: Oklahoma City, OK

The Is No Optimal Portfolio

Post by EDN » Fri Feb 15, 2013 10:16 pm

nisiprius wrote:In theory you are supposed calculate an optimum percentage, based on the standard deviations of the asset classes and the correlations with each other. There's an optimum point on the "efficient frontier." If the standard deviations and correlations don't change, the optimum percentages doesn't change.

But's not supposed to be a matter of gut instinct or taste, it's supposed to be a calculated optimum. Since I don't tilt, I've never looked much into what that calculated optimum is supposed to be. I assume that Bill Schultheis' Coffeehouse Portfolio isn't far off the mark even though he uses round-number percentages.

Of course correlations and standard deviations do change--correlations in particular are wildly unstable. So the mix that was the optimum for the past isn't necessarily the optimum going forward.
I don't think in practice anyone worries about what the optimal mix of small or value is, or if there even is one. You just calculate your required portfolio returns, estimated asset class returns, and appropriate allocation for you. Other considerations include constraints on available indexes, and whether you prefer a stronger exposure to really big and high priced stocks that either do really well (82-99), or very poorly (65-81 and 00-12), or prefer a greater balance of big/small and high/low price (which helped a lot from 65-81, and 00-12 when it was needed most).

There is a 15 year old article in the wiki by Fama Jr that does a great job of showing how this is done. The example US allocation is "moderate" and in the 0.3 range for size and value, using a very unscientific 2/3 large, 1/3 small, and 50/50 market and value (33% S&P 500, 33% large value, 17% micro cap, 17% small value).

Eric

stlutz
Posts: 4660
Joined: Fri Jan 02, 2009 1:08 am

Re: how much tilted is 'a little' or 'a lot'?

Post by stlutz » Fri Feb 15, 2013 10:33 pm

The S&P 600 Value, which is not particularly small or valuey, doesn't represent 2% of the market, you've miscalculated that.
Yes--I rounded up.

"The S&P SmallCap 600 covers approximately 3% of the domestic equities market." In the S&P series, 1/2 of the market cap goes to value, 1/2 to growth. So, the S&P 600 Value index covers 1.5% of the market.

http://www.standardandpoors.com/indices ... --p-us-s--
Anyway, because TSM is dominated by large and mega cap stocks, you need a lot of a small value index that isn't that small or valuey to get healthy size and value exposure.
I actually disagree with nisiprius that this is primarily a mathematical issue; rather, it's what makes logical sense

Suppose a 60 equity/40 bond portfolio.

A "mild" tilt would be appropriate for someone who might say, "I've read about this whole small/valuey thing and just want to try it out for kicks." In that case, about 5% of the portfolio (10% of equities) makes sense. It's enough to have some skin in the game but it's not going to determine the fate of one's portfolio.

A "moderate" tilt would be appropriate for someone who might say, "If I had to bet as to whether value or growth would outperform, I'd say value based on past performance, but I'm also not supremely confident that the future will be like the past." In that case, 15-20% of the portfolio (1/4 - 1/3 of equities) makes logical sense. I would consider the Rick Ferri approach to be a "moderate" tilt.

A "serious" tilt is for someone who might say, 'The professors at the University of Chicago know more about investing than anyone ever has or ever will. The whole Fama/French small/value premium thing has me convinced. There is at least an 85% chance that small/value will beat the market." For that person, the total market should really just be a complement to a small/value dominated portfolio. In that case, SV should be at least 50% of the portfolio.

EDN
Posts: 528
Joined: Sat Dec 29, 2012 7:55 pm
Location: Oklahoma City, OK

S&P 600 and Tilts

Post by EDN » Fri Feb 15, 2013 11:20 pm

stlutz wrote:
Yes--I rounded up.

"The S&P SmallCap 600 covers approximately 3% of the domestic equities market." In the S&P series, 1/2 of the market cap goes to value, 1/2 to growth. So, the S&P 600 Value index covers 1.5% of the market.

http://www.standardandpoors.com/indices ... --p-us-s--

=========================================================================================

I actually disagree with nisiprius that this is primarily a mathematical issue; rather, it's what makes logical sense

Suppose a 60 equity/40 bond portfolio.

A "mild" tilt would be appropriate for someone who might say, "I've read about this whole small/valuey thing and just want to try it out for kicks." In that case, about 5% of the portfolio (10% of equities) makes sense. It's enough to have some skin in the game but it's not going to determine the fate of one's portfolio.

A "moderate" tilt would be appropriate for someone who might say, "If I had to bet as to whether value or growth would outperform, I'd say value based on past performance, but I'm also not supremely confident that the future will be like the past." In that case, 15-20% of the portfolio (1/4 - 1/3 of equities) makes logical sense. I would consider the Rick Ferri approach to be a "moderate" tilt.

A "serious" tilt is for someone who might say, 'The professors at the University of Chicago know more about investing than anyone ever has or ever will. The whole Fama/French small/value premium thing has me convinced. There is at least an 85% chance that small/value will beat the market." For that person, the total market should really just be a complement to a small/value dominated portfolio. In that case, SV should be at least 50% of the portfolio.
S&P 600
The S&P 600 Index holds mid and (mostly) small cap stocks. Small stocks make up approximately the smallest 10% of the market, and small cap indexes with some mid caps included like S&P 600 (similar to MSCI Small Indexes or DFA Targeted Value) represent the smallest 15% of the market. A 50% value sort on the smallest 15% of the market is 7.5%, or 10% once you account for the average drift in stock sizes throughout the year. S&P samples the mid and small cap markets by buying 600 of the eligible 2000 names, so they technically hold 3% or whatever, but it still represents the universe it samples, or about 10%.

Tilts
1. 5% of anything (esp. a small value index that isn't that small or value oriented) isn't mild, its a waste of time. So we can cross that off the list.
2. 75% TSM, 25% S&P 600 Value has only about 0.15 sensitivity to size and value. At the historical 2% and 4% premiums, that is less than a 1% higher-than-market expected return, and a portfolio that doesn't look much different from the market. Probably not worth the effort for most.
3. 40% TSM, 60% S&P 600 Value has about a 0.35 sensitivity to size and value -- a nice balance. About a 2% higher-than-market expected return, which positions you to generate greater wealth, or hold less in stocks (and more in bonds) yet sill earn a similar return as TSM, and provide you some support when TSM goes in the tank (+4% per year higher returns from 2000-2012).

Eric

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

Re: how much tilted is 'a little' or 'a lot'?

Post by pauliec84 » Sat Feb 16, 2013 3:00 am

Tilts
1. 5% of anything (esp. a small value index that isn't that small or value oriented) isn't mild, its a waste of time. So we can cross that off the list.
2. 75% TSM, 25% S&P 600 Value has only about 0.15 sensitivity to size and value. At the historical 2% and 4% premiums, that is less than a 1% higher-than-market expected return, and a portfolio that doesn't look much different from the market. Probably not worth the effort for most.
3. 40% TSM, 60% S&P 600 Value has about a 0.35 sensitivity to size and value -- a nice balance. About a 2% higher-than-market expected return, which positions you to generate greater wealth, or hold less in stocks (and more in bonds) yet sill earn a similar return as TSM, and provide you some support when TSM goes in the tank (+4% per year higher returns from 2000-2012).
One issue is this analysis basically ignores an investor who wants to tilt & be internationally diversified.

An non-DFA investor has practical choices for international small & value.
In fact in my opinion, no feasible small&value choice.

Instead there is (using FF global factors)
EFV:-0.21 SMB, 0.20 HML
PXF: -0.14 SMB, 0.29 HML

VSS: 0.50 SMB, 0.00 HML
PDN: 0.50 SMB, 0.04 SMB
DLS: 0.50 SMB, 0.12 HML

So it is very hard to get a SMB & Value tilt internationally.

User avatar
stevewolfe
Posts: 1529
Joined: Fri Oct 10, 2008 7:07 pm

Re: S&P 600 and Tilts

Post by stevewolfe » Sat Feb 16, 2013 12:17 pm

EDN wrote: Tilts
1. 5% of anything (esp. a small value index that isn't that small or value oriented) isn't mild, its a waste of time. So we can cross that off the list.
Eric
Not true. If we are assuming that we can add a small value tilt to a portfolio to keep an expected rate of return while reducing overall exposure to equities, this isn't a property of a tilt that is lost when you get to a smaller size like 5%.

Consider returns of 3% for bonds, 7% for TSM and 11% for small cap value (using the 4% per year higher returns quoted from 2000-2012). A 5% small value tilt as part of a 35% TSM / 5% small value and 60% bond allocation would have the same return as a 45% TSM / 55% bond allocation.

The difference in equity allocation of 40% vs 45% is small, commensurate with the tilt, but not a waste of time to some.

sdrone
Posts: 108
Joined: Wed Jun 10, 2009 11:17 pm

Re: S&P 600 and Tilts

Post by sdrone » Sat Feb 16, 2013 12:34 pm

EDN wrote:
3. 40% TSM, 60% S&P 600 Value has about a 0.35 sensitivity to size and value -- a nice balance. About a 2% higher-than-market expected return, which positions you to generate greater wealth, or hold less in stocks (and more in bonds) yet sill earn a similar return as TSM, and provide you some support when TSM goes in the tank (+4% per year higher returns from 2000-2012).

Eric
That's interesting, but wouldn't that be small cap equity with a large cap tilt? Heh.

TSM includes something like 9% smallcap + microcap. Ignore midcap for a minute. Your portfolio would be something like 63% to 64% smallcap, the rest large/mid cap.

EDN
Posts: 528
Joined: Sat Dec 29, 2012 7:55 pm
Location: Oklahoma City, OK

More on Tilts

Post by EDN » Sat Feb 16, 2013 1:05 pm

sdrone wrote:
EDN wrote:
3. 40% TSM, 60% S&P 600 Value has about a 0.35 sensitivity to size and value -- a nice balance. About a 2% higher-than-market expected return, which positions you to generate greater wealth, or hold less in stocks (and more in bonds) yet sill earn a similar return as TSM, and provide you some support when TSM goes in the tank (+4% per year higher returns from 2000-2012).

Eric
That's interesting, but wouldn't that be small cap equity with a large cap tilt? Heh.

TSM includes something like 9% smallcap + microcap. Ignore midcap for a minute. Your portfolio would be something like 63% to 64% smallcap, the rest large/mid cap.
It depends on the vehicle(s) you use. I note that swapping S&P 600 Value for Russell 2000 Value requires only a 50/50 split of TSM and SV to get to about the same size and value exposures.

And while we typically say TSM portfolios are "75% large, 15% mid, and 10% small", in reality that "large" category is pretty big, and includes large and mega caps which are of course the behemoths of the market, and I'd guess 45% plus of the TSM portfolio is in these mega-cap names, with the other 30% being in just large. So, yeah, you need a lot of a relatively milder small value index to get you modest size and value exposure.

Incidentally I chose this allocation because it is the closest I could engineer to match a 50% DFA US Core 2, 50% DFA US Vector mix, which is a nice balanced allocation that also holds every stock in the market but is spread about 1/3 large, 1/3 mid, and 1/3 small (and 10% growth, 40% blend, 50% value). So depending on how you get to a particular size and value exposure, I guess it can either look fairly lumpy or perfectly symmetrical.

I'd maybe focus on weighted average market caps, which are about the same for all 3 mixes above (about $40B vs $90B for the market) -- so its definitely smaller, but still quite large.

The mild allocation I came up with just mirrors a 100% DFA US Core 2 allocation.

Like I said, there is no consensus on this, but we can speak in generalities.

Eric

Random Walker
Posts: 2981
Joined: Fri Feb 23, 2007 8:21 pm

Correlations!

Post by Random Walker » Sat Feb 16, 2013 4:39 pm

Everyone focuses on the increased expected returns for small and value tilting. I think the improvement in portfolio efficiency caused by the low correlations between market, value, and small needs to always be appreciated. After all, one needs to look at the portfolio as a whole, rather than thinking about "bets" on risk factors.
From what I remember the correlations are roughly as follows:
Small to market 0.41
Value to market 0.14
Small to Value 0.14

These numbers represent big potential diversification benefits at the portfolio level.

Unfortunately William Bernstein just taught in his short online book that once a great asset class becomes easily investable the diversification benefit goes away. Hopefully only partially goes away.

Dave

stlutz
Posts: 4660
Joined: Fri Jan 02, 2009 1:08 am

Re: how much tilted is 'a little' or 'a lot'?

Post by stlutz » Sat Feb 16, 2013 4:59 pm

From what I remember the correlations are roughly as follows:
Small to market 0.41
Value to market 0.14
Small to Value 0.14
For those who might be new, these are the correlations of the "risk factors" to each other, not the correlations of funds to each other

So, for example, "small to market" is comparing the correlation between 1) going long small cap stocks with 50% of your portfolio and going short large cap stocks with the other 50% and 2) just going long the market.

The correlation of a smallcap value index to the overall market, in contrast, is still quite high. Vanguard lists the correlation of their smallcap value fund to the total market at .95.

dad2000
Posts: 642
Joined: Fri Feb 03, 2012 7:04 pm

Re: how much tilted is 'a little' or 'a lot'?

Post by dad2000 » Sat Feb 16, 2013 5:23 pm

My portfolio (when run through M* Instant XRay) has style values almost identical to to DFA US Core Equity I (DFEOX). I would guess that I'm considered a moderate tilter.

EDN
Posts: 528
Joined: Sat Dec 29, 2012 7:55 pm
Location: Oklahoma City, OK

Core 1 is "Very Mild"

Post by EDN » Sun Feb 17, 2013 11:27 am

dad2000 wrote:My portfolio (when run through M* Instant XRay) has style values almost identical to to DFA US Core Equity I (DFEOX). I would guess that I'm considered a moderate tilter.
dad2000,

I'd actually call that level of tilt (about 0.1 to size and value) to be very mild to non-existent. Double that (Core 2 at about 0.2) would be mild.

DFA created Core 1 as a broad market portfolio with very minimal market tracking error. For those using S&P 500 as their "core" already had some tracking error relative to TSM, so Core 1 simply preserved that level of TE (which wasn't much) by going in the other direction (smaller/more value) but with a greater-than-market expected return (historically only about 0.75% or so).

For those who use Core 1, I cannot remember a case where they aren't augmenting that fund with additional small/value funds or the core-like Vector strategy.

Not saying your mix is wrong, just trying to properly categorize it.

Eric

stlutz
Posts: 4660
Joined: Fri Jan 02, 2009 1:08 am

Re: how much tilted is 'a little' or 'a lot'?

Post by stlutz » Sun Feb 17, 2013 12:10 pm

Hi Eric,

Your comments in this thread are interesting to me. Typically on this board the "tilters" will explain that SV will beat TSM by such a huge amount and with such a high likelihood that one would be dumb not to do it at least to some extent.

Your argument seems to be different. You keep saying that in order for tilting to make any difference you must do it with at least 50% of your equities; <50% = 0%. If the benefits of tilting are so small that they will only be noticed if you do it with most of your equity portfolio, why take the risk of concentrating your investments in 2%/4%/6%/whatever% of the market?

Post Reply