Tilters: how much would you pay for a tilt?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply

How much are you willing to pay over your core (e.g. S&P 500) for a small-cap value tilt?

0.0% - tilting is not worth additional costs
13
19%
0.0-0.1%
10
14%
0.1-0.2%
18
26%
0.2-0.3%
10
14%
0.3-0.4%
7
10%
0.4-0.5%
9
13%
0.5-0.6%
1
1%
0.6-0.7%
2
3%
0.7-0.8%
0
No votes
0.8-0.9%
0
No votes
0.9-1.0%
0
No votes
 
Total votes: 70

Topic Author
DVMResident
Posts: 1814
Joined: Mon Aug 01, 2011 8:15 pm

Tilters: how much would you pay for a tilt?

Post by DVMResident »

This poll has 3 underlying assumption:
(1) the evidence for low-cost investing is strong
(2) the evidence for tilts over the long-term is strong *
(3) you think tilting is a good idea

Tilts funds often have higher ERs than board market funds or their close surrogates (e.g. S&P 500).

401(k) often have limited options and therefore tilting in a 401(k) can be costly. Even if you believe in tilting (I do), there must be some tipping point extra expected returns are exceeded by costs.

How much are willing to for a tilt to small-cap value (probably the most popular tilt on BH) over your core fund (e.g. S&P 500 or Total Stock Market)?
How did you arrive at your answer?

---

* It is understood there can be periods of under-performance. These periods can last a long time.
User avatar
ERMD
Posts: 291
Joined: Thu Dec 26, 2013 10:26 am

Re: Tilters: how much would you pay for a tilt?

Post by ERMD »

in order to tilt, i split domestic equities 50/50 between viiix (s&p index, ER 0.02(!!!) and vieix (ER 0.12). believe it or not, i never gave this discrepancy a lot of thought, since 0.02 is so ridiculously low i feel like i should be paying a few BP more for it anyway. but now i wonder if 10 BP over 30 years is worth the experiment.
between scotch and nothing, i'll take scotch. -- faulkner
berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Tilters: how much would you pay for a tilt?

Post by berntson »

Here is how I think about things. My current target is to have a portfolio with .3 hml and .2 smb. In the US, a portfolio with this sort of tilt has historically resulted in about 200 basis points of annualized outperformance. I figure that there is roughly a 50% chance of things going well. If things go well, I expect to get roughly the historic outperformance (could be higher, could be lower, but the expectation value is the historic value). I figure there is also a 50% chance that things go badly and I get nothing. In which case, I will be out the extra costs of tilting.

This means that the expected value of my tilt is 100 basis points less anything I spend on taxes and fees to get the tilt. I like to build a reasonable margin for error into my portfolio, so I want expenses to be no more than 50 basis points. This leaves in 50 basis points of margin (incase returns are worse than expected in the good case, or in case I have over estimated the likelihood of the good case).

Tilting with Vanguard funds and ETFs is relatively cheap. I figure that I spend about 20 basis points on expense ratios and about another 20 basis points on extra taxes. So I am thus far under by 50 basis point limit.

I think it's worth risking .5% a year of underperformance for a reasonable chance of 2% a year long-term out performance. But not everyone thinks like this.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Tilters: how much would you pay for a tilt?

Post by steve_14 »

berntson wrote:Here is how I think about things. My current target is to have a portfolio with .3 hml and .2 smb. In the US, a portfolio with this sort of tilt has historically resulted in about 200 basis points of annualized outperformance. I figure that there is roughly a 50% chance of things going well. If things go well, I expect to get roughly the historic outperformance (could be higher, could be lower, but the expectation value is the historic value). I figure there is also a 50% chance that things go badly and I get nothing. In which case, I will be out the extra costs of tilting.
So then, 0% chance growth outperforms value and/or large outperforms small over your time horizon?
User avatar
grap0013
Posts: 1892
Joined: Thu Mar 18, 2010 1:24 pm

Re: Tilters: how much would you pay for a tilt?

Post by grap0013 »

I'd pay up to 0.5% US and 0.75% international developed or emerging markets. I arrived at my limits by simple rounding and using the eyeball test to determine what feels about right.

A good SCV EM fund has A LOT higher expected return than something like VWO. I have no regrets about ponying up a little extra for DGS. I also pay more for SFILX which took about 1 day to surpass VEA with a lower SD I might add.
There are no guarantees, only probabilities.
berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Tilters: how much would you pay for a tilt?

Post by berntson »

steve_14 wrote:
berntson wrote:Here is how I think about things. My current target is to have a portfolio with .3 hml and .2 smb. In the US, a portfolio with this sort of tilt has historically resulted in about 200 basis points of annualized outperformance. I figure that there is roughly a 50% chance of things going well. If things go well, I expect to get roughly the historic outperformance (could be higher, could be lower, but the expectation value is the historic value). I figure there is also a 50% chance that things go badly and I get nothing. In which case, I will be out the extra costs of tilting.
So then, 0% chance growth outperforms value and/or large outperforms small over your time horizon?
That was just a back of the envelope estimate. I judge the odds of large and growth beating small and value over the long-term and across international markets to be roughly equal to the odds of small and value beating large and growth by, say, double the long-term average. So the two more or less cancel out and we can ignore them when making simple cost estimates.

Imagine a normal distribution centered on 1%. That's what I take to be a good estimate of future small and value outperformance for a moderately tilted portfolio like my own. Sure, the left tail has small and value underperforming, but the the right tail has it significantly over performing.
User avatar
sans souliers
Posts: 178
Joined: Sun Apr 08, 2012 1:38 pm
Location: By the Quinnipiac

Re: Tilters: how much would you pay for a tilt?

Post by sans souliers »

I have to admit, I'm not sure what a tilt is, so I voted -0-

I stay away from anything I don't understand clearly. Even poll questions.
Sometimes pessimism leaves me pretty well prepared for when things don't go my way, and pleasantly surprised when they do.
User avatar
grap0013
Posts: 1892
Joined: Thu Mar 18, 2010 1:24 pm

Re: Tilters: how much would you pay for a tilt?

Post by grap0013 »

berntson wrote: That was just a back of the envelope estimate. I judge the odds of large and growth beating small and value over the long-term and across international markets to be roughly equal to the odds of small and value beating large and growth by, say, double the long-term average. So the two more or less cancel out and we can ignore them when making simple cost estimates.

Imagine a normal distribution centered on 1%. That's what I take to be a good estimate of future small and value outperformance for a moderately tilted portfolio like my own. Sure, the left tail has small and value underperforming, but the the right tail has it significantly over performing.
What kinds of estimates do you use for odds of TSM outperforming bonds? Hint: they are similar to SCV over TSM both in magnitude and consistency. Long term the trend looks good: http://www.multifactorworld.com/Lists/P ... aspx?ID=95

Here's another good one too: http://us.dimensional.com/pdf/the_risk_ ... eliver.pdf
There are no guarantees, only probabilities.
KlangFool
Posts: 31526
Joined: Sat Oct 11, 2008 12:35 pm

Re: Tilters: how much would you pay for a tilt?

Post by KlangFool »

TS,

Expense ratio

VTSAX - 0.05%
DFSVX - 0.52% in my 401K

Difference = 0.47%

KlangFool
30% VWENX | 16% VFWAX/VTIAX | 14.5% VTSAX | 19.5% VBTLX | 10% VSIAX/VTMSX/VSMAX | 10% VSIGX| 30% Wellington 50% 3-funds 20% Mini-Larry
berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Tilters: how much would you pay for a tilt?

Post by berntson »

grap0013 wrote:
berntson wrote: That was just a back of the envelope estimate. I judge the odds of large and growth beating small and value over the long-term and across international markets to be roughly equal to the odds of small and value beating large and growth by, say, double the long-term average. So the two more or less cancel out and we can ignore them when making simple cost estimates.

Imagine a normal distribution centered on 1%. That's what I take to be a good estimate of future small and value outperformance for a moderately tilted portfolio like my own. Sure, the left tail has small and value underperforming, but the the right tail has it significantly over performing.
What kinds of estimates do you use for odds of TSM outperforming bonds? Hint: they are similar to SCV over TSM both in magnitude and consistency. Long term the trend looks good: http://www.multifactorworld.com/Lists/P ... aspx?ID=95

Here's another good one too: http://us.dimensional.com/pdf/the_risk_ ... eliver.pdf
Thanks grap! Those links are really helpful, especially the DFA link. I agree that when combined, the reliability of the historic small and value premiums is quite impressive, especially in the US. I recently put together a spreadsheet for backtesting the US factor returns. Between 1940 and 1990, there were no rolling five-year periods in which the overall premium was negative (to any significant extent) for a moderately tilted portfolio like mine or Robert T's.

If the future looks anything like the historical data, I'm not worried. I'm mostly worried about data snooping, errors in the historical data, and the possibility that whatever premium there was will be arbitraged away. Even in the US historical data, there is evidence that premiums are slowly shrinking. If we draw a trendline through the premium for a modestly tilted portfolio, we get this:

Image

(The numbers on the x axis are years since 1927.)

My own (somewhat uninformed) view is that much of the small premium was, historically, a liquidity premium. As those markets have become more liquid, that premium has started evaporating. But there is more to the small and value premiums than just liquidity, so we shouldn't expect them to disappear entirely, even in the long run.

It is also interesting to note that the the small and value premiums have been quite robust even after Fama and French published their common risk factors paper in 1992. The markets have known about the premiums for twenty years and they still exist. If they were easy to arbitrage away, it likely would have happened by now. On the other hand, the vast m
User avatar
grabiner
Advisory Board
Posts: 35307
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Tilters: how much would you pay for a tilt?

Post by grabiner »

I am willing to pay a small amount for a tilt, because I believe the small-cap and value premium is primarily a risk premium. I made the decision not to get my value tilt in my taxable account, so I have always held only as much value as I could fit in my IRA and in those retirement plans with a value fund option. With the development of ETFs and qualified dividends becoming permanent, I would now be willing to get a value tilt in my taxable account, but I no longer need to.

I am willing to pay more for a diversification benefit. I hold the iShares emerging markets small-cap index (0.67% expense) rather than holding all of my international in Vanguard's international small-cap (0.20%, would presumably be about 0.30% for the emerging portion) because I expect emerging-markets small-cap to be a good diversifier, but my target weight to emerging markets small-cap is still only half what it would be if cost weren't an issue. Similarly, I bought Vanguard's Global Real Estate Index in my Roth IRA before it offered Admiral shares.
Wiki David Grabiner
User avatar
grap0013
Posts: 1892
Joined: Thu Mar 18, 2010 1:24 pm

Re: Tilters: how much would you pay for a tilt?

Post by grap0013 »

grabiner wrote:I am willing to pay more for a diversification benefit.
I think this is an excellent point. Even if the tilts do not outperform over the long haul, there is still great value when they are zigging while other holdings are zagging. It helps smooth out both short and long term returns.
There are no guarantees, only probabilities.
Topic Author
DVMResident
Posts: 1814
Joined: Mon Aug 01, 2011 8:15 pm

Re: Tilters: how much would you pay for a tilt?

Post by DVMResident »

KlangFool wrote:TS,

Expense ratio

VTSAX - 0.05%
DFSVX - 0.52% in my 401K

Difference = 0.47%

KlangFool
Exactly the point of the poll is to help decide when the tilt is too expensive.

It is easy to compare VG funds in the abstract, but most of us has the bulk of investments in 40_(_) plans, where the options limited.
lazyday
Posts: 3849
Joined: Wed Mar 14, 2007 10:27 pm

Re: Tilters: how much would you pay for a tilt?

Post by lazyday »

The more a fund is tilted, the less of it needed and the higher the ER a tilter should be willing to pay.

A "Pure" small value fund might have a high ER but still be worth it for a tilter.
Streptococcus
Posts: 425
Joined: Wed Jan 02, 2013 11:17 pm

Re: Tilters: how much would you pay for a tilt?

Post by Streptococcus »

I am not willing to pay more than 0.25% in ER.
My tilt:
VBR 0.1%
VSS 0.2%
Vemax 0.15%

My total funds are cheaper :)
User avatar
grabiner
Advisory Board
Posts: 35307
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Tilters: how much would you pay for a tilt?

Post by grabiner »

grap0013 wrote:
grabiner wrote:I am willing to pay more for a diversification benefit.
I think this is an excellent point. Even if the tilts do not outperform over the long haul, there is still great value when they are zigging while other holdings are zagging. It helps smooth out both short and long term returns.
Tilting to value doesn't improve diversification; it's just as likely to lose when growth outperforms value as to gain when value outperforms growth. Similarly for US small-cap, at least in a US-centered portfolio.

In contrast, I would expect that overweighting emerging markets, international small-cap, and international real estate does improve diversification, as these asset classes are less likely to be correlated with the rest of a portfolio.
Wiki David Grabiner
KlangFool
Posts: 31526
Joined: Sat Oct 11, 2008 12:35 pm

Re: Tilters: how much would you pay for a tilt?

Post by KlangFool »

DVMResident wrote:
KlangFool wrote:TS,

Expense ratio

VTSAX - 0.05%
DFSVX - 0.52% in my 401K

Difference = 0.47%

KlangFool
Exactly the point of the poll is to help decide when the tilt is too expensive.

It is easy to compare VG funds in the abstract, but most of us has the bulk of investments in 40_(_) plans, where the options limited.
Hi,

I do not understand your post.

1) My portfolio is approximately 50% taxable and 50% tax advantage A/C (401K / IRAs).

2) In my 401K A/C, I have the option to buy DFSVX with expense ratio of 0.52%. I have the choice of buying DFSVX or allocate that portion to VTSAX at Taxable A/C with expense ratio of 0.05%.. I believe it is worthwhile to do a small value tilt at extra expense ratio of 0.47%. Hence, I buy DFSVX at my 401K A/C.

I believe I answered the TS's question. There is nothing abstract in my response. This is exactly what I do in my portfolio.

KlangFool
30% VWENX | 16% VFWAX/VTIAX | 14.5% VTSAX | 19.5% VBTLX | 10% VSIAX/VTMSX/VSMAX | 10% VSIGX| 30% Wellington 50% 3-funds 20% Mini-Larry
berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Tilters: how much would you pay for a tilt?

Post by berntson »

grabiner wrote:
grap0013 wrote:
grabiner wrote:I am willing to pay more for a diversification benefit.
I think this is an excellent point. Even if the tilts do not outperform over the long haul, there is still great value when they are zigging while other holdings are zagging. It helps smooth out both short and long term returns.
Tilting to value doesn't improve diversification; it's just as likely to lose when growth outperforms value as to gain when value outperforms growth. Similarly for US small-cap, at least in a US-centered portfolio.

In contrast, I would expect that overweighting emerging markets, international small-cap, and international real estate does improve diversification, as these asset classes are less likely to be correlated with the rest of a portfolio.
If long-term historic returns are a good guide to the future, then there will be small and value premiums that are positive and not entirely correlated with the market premium. This would let us build a tilted portfolio with less risk and the same returns as a non-tilted portfolio, or one with higher expected returns and the same risk. This is just the familiar trade-off between risk and returns.
User avatar
grap0013
Posts: 1892
Joined: Thu Mar 18, 2010 1:24 pm

Re: Tilters: how much would you pay for a tilt?

Post by grap0013 »

grabiner wrote: Tilting to value doesn't improve diversification; it's just as likely to lose when growth outperforms value as to gain when value outperforms growth. Similarly for US small-cap, at least in a US-centered portfolio.

In contrast, I would expect that overweighting emerging markets, international small-cap, and international real estate does improve diversification, as these asset classes are less likely to be correlated with the rest of a portfolio.
US value and small correlations to the overall US market have been ~0.1 and ~0.4 in the past. I assume international small or value in relation to the US market correlations are even smaller. However, 0.1 and 0.4 are very good. They can add value when beta has poor returns. Just take 1966-1982 for example. Equity premium was basically nil, but investors who had small and/or value allocations had much better returns. So it's not so much about something beating something else, but rather, low correlation assets may help bridge some troughs on the growth of your portifolio.
There are no guarantees, only probabilities.
RNJ
Posts: 863
Joined: Mon Apr 08, 2013 9:06 am

Re: Tilters: how much would you pay for a tilt?

Post by RNJ »

Is it possible that, as with so many other aspects of portfolio construction, it is more helpful to look at expense ratios from the perspective of the portfolio-as-a-whole?

With respect to tilting, I've picked my spots. I own sizable percentages in ILV (EFV, er .4% and PXF, er .47%), ex-US Small (VSS er .2%) and US Small Value (VIOV er .2% and PXSV er .39%).

The er on the portfolio as a whole is .17% - a number with which I am comfortable.
User avatar
grabiner
Advisory Board
Posts: 35307
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Tilters: how much would you pay for a tilt?

Post by grabiner »

RNJ wrote:Is it possible that, as with so many other aspects of portfolio construction, it is more helpful to look at expense ratios from the perspective of the portfolio-as-a-whole?
The reason to look specifically at the cost of the tilt is that both the benefit and the cost depend on the fraction of your portfolio you tilt, so the cost determines whether the tilt is worthwhile. It costs twice as much to put half of an 80%-stock portfolio into small value as to put half of a 40%-stock portfolio into small value, but the 80%-stock portfolio gains the diversification benefit over twice as much stock.

In my own portfolio, the cost of small-cap value tilting is trivial (Admiral shares of Small-Cap Value Index), while the cost of emerging markets small-cap tilting is very high (iShares EEMS, and even the large-cap Emerging Markets Index is significantly more costly because of lower qualified dividends), so I do a full-weight tilt in my US stock, and a reduced-weight tilt to emerging markets and foreign small-cap.
Wiki David Grabiner
User avatar
zaboomafoozarg
Posts: 2431
Joined: Sun Jun 12, 2011 12:34 pm

Re: Tilters: how much would you pay for a tilt?

Post by zaboomafoozarg »

I just ran into a situation that reminded me of this thread. My new job has a 401k which doesn't offer SCV, but does have a total US market fund for a whopping .02% ER.

That's .35% cheaper than the DFA SCV fund I currently use in my old 401k... and I only voted .2% to .3% on this poll. Decisions, decisions. :D

I suppose I could roll the old 401k to an IRA and use VBR for .09%. It would mess up in-service after-tax 401k rollovers though.
Post Reply