"Factor investing"

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Taylor Larimore
Advisory Board
Posts: 27606
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

"Factor investing"

Post by Taylor Larimore » Thu Sep 08, 2016 10:02 pm

Bogleheads:

In recent years, the investment industry has begun talking about "factors" as a way to beat the market. According to this MSCI paper, there are six "factors": Value, Low size, Low volatility, High yield, Quality and Momentum (I wonder why "Low-cost" was not included). According to MSCI (which provides data for the financial industry): "these factors are grounded in academic research and have solid explanations as to why they historically have provided a premium."

I am skeptical, primarily because nearly all academic research about investing is based on past performance which experienced investors and the government warn us against.

I have also learned from experience (and professor Burton Malkiel) who wrote: "The very popularity of any investment style will sow the seeds of its own destruction." I well remember the industry-promoted popularity of Penny stocks, Day Trading, the "Nifty-Fifty," Dogs of the Dow, IPOs, Commodities, etc..

This 2003 post of mine, containing an article, PREDICTING THE PAST, recalls the "factors" of earlier days.

I am confident that every mutual fund manager knows about "factors" and many managers use them. Nevertheless, the majority of fund managers underperform simple broad market index funds.

A study by Rick Ferri & Alex Benke found that an all index 3-fund market portfolio outperformed active portfolios 82.9% of the time during a 16-year period (1997-2012). That's good enough for me. It also reflects results from Three-Fund Portfolio forum topic and my book, The Bogleheads' Guide to the Three-Fund Portfolio.
"By and large I do not approve of factor funds." Jack Bogle in Rick Ferri podcast
"You don’t need to constantly add new asset classes or investments just because investment firms keep bringing them out. In fact, if you do, you’re more likely to end up with an unwieldy hodge-podge of investments that’s difficult to manage rather than a simpler portfolio that more efficiently balances risk and return." -- Walter Updegrave
"Profitable strategies, if they exist at all, do not last for very long. As soon as they are discovered, they are acted upon by the investment community bidding up the price of the relevant assets, thus eliminating their excess return." -- Bill Bernstein in The Intelligent Asset Allocator
The oldest multi-factor fund I could find is PNC Multi-Factor Small-Cap Value C (PSVCX). It's 15-year average return on 7/06/2018 was 7.12%. Meanwhile, Total Stock Market (VTSAX) 15-year average return was 9.80%.

AQR is noted for its factor funds. This 2017 study by Tom Allen and Mark Hebner found: (16 funds) have underperformed their respective benchmarks since inception and only 8 funds outperformed.
Wall Street has a particular fetish for inventing unnecessarily complicated ways of achieving simple goals. Most other professions, so far as I can tell, are more comfortable offering simple ways to do simple things. On Wall Street, complicated ways to do simple things have bigger payoffs — for Wall Street, that is. So complexity proliferates, and it gets encrusted with jargon to impress investors into thinking there’s something special about it. -- Jason Zweig, author and Wall Street Journal columnist
Vanguard: Factor-based funds are a form of active management. They offer the potential to achieve specific risk and return objectives by purposely and explicitly "tilting" portfolios toward certain stock characteristics, like recent momentum, higher quality, or lower stock prices.

But they come with significantly more risk than you'd experience investing in the broader stock market.

Smart Beta Is Making This Strategist Sick

Factor Investing. What Went Wrong?

Best wishes.
Taylor
Last edited by Taylor Larimore on Sat Sep 15, 2018 2:47 pm, edited 18 times in total.
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
nedsaid
Posts: 10644
Joined: Fri Nov 23, 2012 12:33 pm

Re: "Factor investing"

Post by nedsaid » Thu Sep 08, 2016 10:12 pm

Taylor, thank you for posting this. In a dividend thread, I cited this very paper in making the argument that dividends are important as a factor in stock returns. Larry Swedroe pretty much dismissed this and pretty much said this (dividends as a factor) was not supported by any academic research. I thought this paper looked awfully familiar. Good to see that someone else saw the MSCI paper and linked to it.

I like Larry a whole lot but it did irk me a bit to be dismissed. Hopefully others on the thread will look at this white paper and draw their own conclusions.

I like the chart that shows factor investing in a continuum between passive and active.

As far as my opinion on factor investing versus market cap weighted indexing, I could go either way. Those who want to invest in 3-4 of the broad index funds will get no argument from me, that is a particularly good way to invest because you get very broad diversification at a very low cost.

I do like the idea of investing across factors, as Larry Swedroe recommends. Different factors perform well at different times and this should decrease volatility in a portfolio. Us old timers remember when these used to be called investment styles.
A fool and his money are good for business.

User avatar
FIREchief
Posts: 2741
Joined: Fri Aug 19, 2016 6:40 pm

Re: "Factor investing"

Post by FIREchief » Thu Sep 08, 2016 11:22 pm

Taylor Larimore wrote:
Beware of "factors" that claim they will beat the market.

Best wishes.
Taylor
I think this pretty much says it all. I guess maybe we could change it to "Beware of "<anything>" that claim they will beat the market." :sharebeer
I am not a lawyer, accountant or financial advisor. Any advice or suggestions that I may provide shall be considered for entertainment purposes only.

User avatar
FIREchief
Posts: 2741
Joined: Fri Aug 19, 2016 6:40 pm

Re: "Factor investing"

Post by FIREchief » Thu Sep 08, 2016 11:26 pm

nedsaid wrote: not supported by any academic research
This is where he usually loses me. The very term "academic research" usually conjures up negative thoughts. :annoyed
I am not a lawyer, accountant or financial advisor. Any advice or suggestions that I may provide shall be considered for entertainment purposes only.

User avatar
snowshoes
Posts: 348
Joined: Tue Feb 03, 2015 12:33 pm

Re: "Factor investing"

Post by snowshoes » Thu Sep 08, 2016 11:47 pm

Taylor Larimore wrote:Bogleheads:
there are six "factors": Value, Low size, Low volatility, High yield, Quality and Momentum. According to MSCI (which provides data for the financial industry): "these factors are grounded in academic research and have solid explanations as to why they historically have provided a premium."
Best wishes.
Taylor
I'm recalling the recent 'smart beta' claims, what happened to that investment product or strategy selling fad? This stuff must be to reel in market newbies commissions, fees, costs, etc. Great catch Taylor!

User avatar
PeteyDink
Posts: 86
Joined: Fri May 06, 2016 5:18 pm

Re: "Factor investing"

Post by PeteyDink » Fri Sep 09, 2016 5:12 am

It's not about trying to beat the market, but investing in an asset class that provides more return for more risk.

DayOfChange
Posts: 61
Joined: Tue Jun 14, 2016 5:37 pm

Re: "Factor investing"

Post by DayOfChange » Fri Sep 09, 2016 6:30 am

i have also learned from experience (and professor Burton Malkiel) who wrote: "The very popularity of any investment style will sow the seeds of its own destruction." I well remember the popularity of Penny stocks, Day Trading, the "Nifty-Fifty," Dogs of the Dow, IPOs, Commodities, etc..
Are index funds immune to this phenomenon? They seem to be getting wildly popular

runnerguy
Posts: 52
Joined: Tue Dec 01, 2015 9:09 am

Re: "Factor investing"

Post by runnerguy » Fri Sep 09, 2016 7:12 am

Like
DayOfChange wrote:
i have also learned from experience (and professor Burton Malkiel) who wrote: "The very popularity of any investment style will sow the seeds of its own destruction." I well remember the popularity of Penny stocks, Day Trading, the "Nifty-Fifty," Dogs of the Dow, IPOs, Commodities, etc..
Are index funds immune to this phenomenon? They seem to be getting wildly popular

Da5id
Posts: 2061
Joined: Fri Feb 26, 2016 8:20 am

Re: "Factor investing"

Post by Da5id » Fri Sep 09, 2016 7:29 am

DayOfChange wrote:
i have also learned from experience (and professor Burton Malkiel) who wrote: "The very popularity of any investment style will sow the seeds of its own destruction." I well remember the popularity of Penny stocks, Day Trading, the "Nifty-Fifty," Dogs of the Dow, IPOs, Commodities, etc..
Are index funds immune to this phenomenon? They seem to be getting wildly popular
Presumably any sector index fund is subject to it. e.g. folks chasing yields pouring into REIT indexes, high dividend indexes, etc. The broader the index you purchase, it appears to me less likely you will be subject to the risk of your investment style fad going bust. Many of the fads Malkiel lists are concentrations into a style or sector. You could I guess make the case that broad US index funds (rather than global) are subject to the risk too, who knows the US could be the next Japan...

But if you mean index vs active with a similar investment scope, I doubt it is a risk I'd lose sleep over :)

runnerguy
Posts: 52
Joined: Tue Dec 01, 2015 9:09 am

Re: "Factor investing"

Post by runnerguy » Fri Sep 09, 2016 7:33 am

In response to the OPs link to 2003 about the efficient frontier... The efficient frontier 'works' (even after 2002) if implemented and systematically updated. Nearly all investors who use efficient frontier don't use it in that manner. They simply pick a buy and hold portfolio based on what the efficient frontier is currently. As the efficient frontier changes an investor needs to radically change their portfolio regularly which requires high (er) turnover. Investors who would monthly or quarterly update their portfolio to the efficient frontier would get the intended risk adjusted returns.

It is similar to current momentum/volatility weighted strategies or adaptive asset allocations.

User avatar
nisiprius
Advisory Board
Posts: 37048
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: "Factor investing"

Post by nisiprius » Fri Sep 09, 2016 7:34 am

PeteyDink wrote:It's not about trying to beat the market, but investing in an asset class that provides more return for more risk.
It can sometimes be hard to make out exactly what factor advocates are claiming. Because, of course, there is no trick to getting more return for more risk. You just increase your stock allocation.

If Vanguard LifeStrategy Moderate Growth, 60/40, blue line, isn't risky enough for you, then goose it up to Vanguard LifeStrategy Growth, 80/20, orange line. Higher highs. Lower lows. 7.87% since inception, versus 7.59% for Moderate. More stocks = more return for more risk. All you need to do is look, it's obvious, no Sharpe ratios or factor loadings needed. (This works all the way up to 100% stocks. If that isn't risky enough for you, then, yes, you need to do something different).

Source
Image

So factor investing can't be just about more return for more risk. Factor advocates seem to make two claims, both squishy.

The first one is squishy because I can't even figure out whether factor advocates are actually making it. That claim is that their preferred asset classes offer not just higher risk and higher return, but higher risk-adjusted return than the market. Sometimes they seem to be saying this, sometimes not. For example, I once quoted Rolf Banz' 1981 words, "It is found that smaller firms have had higher risk-adjusted returns, on average, than larger firms. This 'size effect' has been in effect for at least forty years and is evidence that the capital asset pricing model is mis-specified." I took that to be a claim that the asset class intrinsically has higher risk-adjusted return than the market, but a factor advocate told me that I was wrong, Banz was simply pointing out a curiosity over a specific period of time with no implication that he thought it would persist.

The existence of an asset class with intrinsic, persistent, higher risk and higher return than the market doesn't violate the efficient market hypothesis, but one that has intrinsic, persistent, higher risk-adjusted turn does. Perhaps that's one reason why the claim is squishy. At times factor advocates seem to be saying that the various "premia" are explained by risk. At other times, we are told that they must have a behavioral component, and that although it is hard to understand why they persist, empirically they have persisted for a long time... and that they might go away at any time when the efficient market wises up, but nevertheless we should plan our portfolios on an assumed likelihood that they won't.

That is, this argument for factor-based investing is based on personal judgement that are certain irrational anomalies will persist even after everyone has heard about them. Like me and the giant box of Nature Valley Almond Butter Biscuits my wife brought home from Costco, we can't stop ourselves.

The second claim is that even if the exotic asset classes do not actually have higher risk-adjusted return than the market, nevertheless, if they are uncorrelated, a portfolio of them can and does have higher risk-adjusted return than the market, and that this is an effect which cannot be arbitraged away because it is a characteristic of the portfolio rather than of anything specific thing within it.

This is squishy because it has a dual aspect. It is just math that if you have a set of real or hypothetical data for such a portfolio over a long period of time, and if, over that time period, the asset classes had the right relationships (which are much more stringent than just low correlation), the portfolio will in fact have had a calculated risk-adjusted reward that's higher than the (weighted) sum-of-its-parts, by amounts in the ballpark of 0.5%, 1%, even 1.5%/year.

There are, however, two problems with this: lack of persistence, and rarity of existence.

Persistence: extra risk-adjusted return emerges from the interplay of characteristics and relationships between the asset classes that may not be any more persistent than return itself. Bill Miller beat the S&P 500 in return for fifteen years in a row, but that return did not persist. "Commodities" (using whatever dataset PortfolioVisualizer uses) has a U.S. market correlation from 1972 to 2006 of -0.17, and improved the risk-adjusted-return hypothetical backtested portfolios enough to convince fund companies to start including them in target-date funds, but that correlation did not persist; for 2007 through 2015, it was 0.62.

Rarity: The second is that it is not enough for two assets to have low correlation. In order to see a meaningful benefit, they have to have a) genuinely low correlation, not just "less than 1," and the big magic occurs only when you get to negative correlation; b) roughly comparable volatility; cash has zero correlation with stocks, but does not improve (nor harm) the risk-adjusted return of stocks because it has zero volatility; c) most important, and most often lacking, both assets need to have a decent and roughly comparable return.

It is very hard for a lousy investment to improve a portfolio. It's theoretically possible for a zero-return investment to improve a portfolio if it has a negative correlation with stocks over the time period examined; such an asset has the amazing characteristic of erasing or cancelling out some stock risk, without incurrent a cost in the sense of an actual loss. This can occur by happenstance over limited periods of time, but I think over long periods it's the investing equivalent of perpetual motion.

So, in real life, what you see is that if, in the past, some factor-based portfolios was made of asset classes with the requisite rare characteristics, then, in the past, they did have higher risk-adjusted reward. The problem is that the requisite characteristics, in a form strong enough to matter, are rare, and often not persistent.

(Of late, factor advocates have shifted from claims of investing in uncorrelated asset classes to language that talks about investing in uncorrelated risk factors. Risk factors are uncorrelated because that's how they're mathematically constructed. You have to be very careful in listening, because you can't invest in a risk factor itself, you can only invest in assets. Of late, the idea has been to use long-short portfolios to cancel out the market factor and allow the factors to emerge in a purer form, and then leverage the factors up to make them strong enough to have a noticeable effect. The problem is that although the factors themselves do have the requisite low correlation, pure factors don't have much return. The long-short portfolios that try to cancel out the market factor are also cancelling out a lot of return at the same time. Another potential problem is that this kind of strategy depends on the factor loadings and correlations of actual assets being stable. If the assets in the portfolio suddenly start to have big changes in the factors they contain, the leverage and short positions are going to magnify that difference.)
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.

Levett
Posts: 4177
Joined: Fri Feb 23, 2007 2:10 pm
Location: upper Midwest

Re: "Factor investing"

Post by Levett » Fri Sep 09, 2016 7:59 am

Hi Taylor,

Thank you for bringing back the memory of Ozark--a man for all seasons!

I must say that the "factor" I give greatest weight to as a retiree is (drum roll :wink: ): satisfaction.

Lev

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Fri Sep 09, 2016 8:38 am

One key thought here re past performance isn't predictive. That refers to the performance of ACTIVE MANAGERS, not factors.

Now the problem we have is that there are over 600 factors in the "factor zoo." The question is how to decide which ones to consider investing in.

IMO there are good criteria that one can establish, the same criteria we use for market beta as a factor.
The factor should provide ADDITIONAL explanatory power in the cross-section of returns and provide a premium that is
Persistent across economic regimes over long periods
Pervasive across industries, sectors, countries, regions, and even asset classes
Robust to various definitions to address issue of data mining
Intuitive explanation for why it should persist (either risk based or behavioral)
Implementable, meaning survives transactions costs
And isn't subsumed by other factors


As I explain and present the evidence on each of these issues there are a small number considering (market beta, size, value, momentum, quality/profitability, carry and term. That's what my new book will cover, presenting the evidence on each of the criteria so each person can decide for themselves, making informed decisions. Note we just finished the editing this morning and am told the book should be out by November. Skepticism is of course healthy, and it is what allows us to reject the vast majority of factors that have been "discovered" and written up. But there are also factors with strong empirical support for believing why you should EXPECT to have a premium.

I would add that if you don't invest in factors other than market beta than you have all your eggs in that one risk basket, a basket which has the risk of performing very poorly for a very long time. Personally I would rather have a portfolio that is diversified across different sources of risk/return, where there is strong evidence that one should expect an ex-ante premium.

Nedsaid,
RE dividends, to be clear, what I have said is a high dividend strategy does produce a premium, it's just that it's relatively weak in terms of statistical evidence (lower t-stats) and it's lower premium and bottom line it's explanatory power is subsumed by other factors that do a better job of explaining returns. And as noted dividends themselves should in theory have no explanatory power, independent of other factors. There is simply no logic or theory to support it--at least IMO.

am
Posts: 2794
Joined: Sun Sep 30, 2007 9:55 am

Re: "Factor investing"

Post by am » Fri Sep 09, 2016 9:08 am

I have a hard time believing that when beta performs poorly for a long time, that the other factors will come to the rescue And give me positive returns. Additionally, there are issues with fund availability and costs across workplace accounts, iras, etc. Hard to believe that a little value, momentum, etc on top of my base in total market beta heavy funds will make a big difference. Also, seems like beta risk has paid of well in the past just like all the back tested factors. Having said that, is there a low cost indexed all factor fund out there from the likes of Vanguard or Fidelity?

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

Re: "Factor investing"

Post by in_reality » Fri Sep 09, 2016 9:19 am

am wrote: Having said that, is there a low cost indexed all factor fund out there from the likes of Vanguard or Fidelity?
All factor?
US domiciled?
From Vanguard?

No.

Through the UK though, there are the Vanguard Global Value Factor ETF, Vanguard Global Momentum Factor ETF, Vanguard Global Liquidity FactorETF (shares that are less frequently traded) and Vanguard Global Minimum Volatility ETF. 0.22ER

Call_Me_Op
Posts: 7063
Joined: Mon Sep 07, 2009 2:57 pm
Location: Milky Way

Re: "Factor investing"

Post by Call_Me_Op » Fri Sep 09, 2016 9:43 am

Taylor Larimore wrote: Beware of "factors" that claim they will beat the market.
There is a lot of hand-waving on both sides of this issue. What is meant by "beat the market?' Does this mean beat it in absolute terms or on a risk adjusted basis? I think it is quite reasonable to expect various asset sub-classes to "beat the market" on an absolute basis, since these sub-classes (e.g., small-cap stocks) are riskier than the cap-weighted market index. I define risk in a very specific (and unconventional way) in conjunction with this claim. Risk refers to security behavior that frightens investors.
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein

pkcrafter
Posts: 13126
Joined: Sun Mar 04, 2007 12:19 pm
Location: CA
Contact:

Re: "Factor investing"

Post by pkcrafter » Fri Sep 09, 2016 9:54 am

Taylor, thank you for being the Bogleheads' unfailing ballast. I remember Ozark and I also remember when you made the post. I have successfully followed your simple strategy since 1998. Simplified, diversified, efficient portfolio = efficient investor. Don't fix it if it ain't broke.

runnerguy wrote:
As the efficient frontier changes an investor needs to radically change their portfolio regularly which requires high (er) turnover. Investors who would monthly or quarterly update their portfolio to the efficient frontier would get the intended risk adjusted returns.

It is similar to current momentum/volatility weighted strategies or adaptive asset allocations.

You may follow the idea of trying to keep up with a constantly changing EF, but be aware that this is not a Boglehead strategy.

Paul
When times are good, investors tend to forget about risk and focus on opportunity. When times are bad, investors tend to forget about opportunity and focus on risk.

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Fri Sep 09, 2016 10:01 am

am
Don't need the factors to provide positive returns even to add value, just lower losses. And note that MOM is negatively correlated with beta and time series MOM provides tail risk hedging properties.
Also consider the performance of value say from 2000-02 when markets crashed and value premium was perhaps largest ever.
Also profitability and quality factors tend to perform better in negative markets, providing diversification benefits as well
Best wishes
Larry

User avatar
bertilak
Posts: 6157
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Re: "Factor investing"

Post by bertilak » Fri Sep 09, 2016 10:55 am

nisiprius wrote:
PeteyDink wrote:It's not about trying to beat the market, but investing in an asset class that provides more return for more risk.
It can sometimes be hard to make out exactly what factor advocates are claiming. ...
Nisi,

Thanks for that post. It has put a little meat behind my vague misgivings about factor investing.
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker, the Cowboy Poet

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Fri Sep 09, 2016 11:01 am

t's not about trying to beat the market, but investing in an asset class that provides more return for more risk.
Or premium return for a DIFFERENT type of risk, not more but different, which is important
Larry

User avatar
Kevin M
Posts: 10286
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: "Factor investing"

Post by Kevin M » Fri Sep 09, 2016 11:25 am

Taylor Larimore wrote: In recent years, the investment industry has begun talking about "factors" as a way to beat the market <snip>(I wonder why "Low-cost" was not included).

I think the reason is that most of the research that identifies factors examines returns of stocks, not of funds. Some research does "factor" in fund expenses though, for example the research that finds that factors explain most of the performance of active managers after costs; from the paper you shared the link to:

For instance, Fama and French (2010) found that mutual funds in aggregate (CRSP database) experienced net returns that underperformed the Fama-French factor benchmarks by about the costs in expense ratios from 1984 to 2006

Taylor Larimore wrote:According to MSCI (which provides data for the financial industry): "these factors are grounded in academic research and have solid explanations as to why they historically have provided a premium."

I am skeptical, primarily because nearly all academic research about investing is based on past performance which experienced investors and the government warn us against.

I am skeptical too, but for the same reason I'm skeptical that stocks (factor = beta) will outperform bonds (factors = term and credit) over any particular period of interest to me, no matter how long.

It's also important to point out that good research examines out of sample data after establishing an hypothesis about a factor. Larry has articulated the criteria for establishing the merit of any particular factor.

Taylor Larimore wrote:I have also learned from experience (and professor Burton Malkiel) who wrote: "The very popularity of any investment style will sow the seeds of its own destruction."

Another good caveat, yet we see the continued outperformance, for example, of small-value stocks despite many years of the small and value factors being well known. Just comparing Vanguard small-cap value (VISVX, not necessarily the best for loading on small-value, but good enough for me) and Vanguard Total Stock Market (VTSMX) for the longest period possible, June 1998 - August 2016, we see (Backtest Portfolio Asset Allocation):

Higher return? Yes: VISVX 8.53%, VTSMX 6.15%

Riskier? Yes, with these standard deviations: VISVX 19.38%, VTSMX 16.09%

Higher risk-adjusted return? Yes, with these Sharpe ratios: VISVX 0.43, VTSMX 0.34

Taylor Larimore wrote:I am confident that every mutual fund manager knows about "factors" and many managers use them. Nevertheless, the majority of fund managers underperform simple broad market index funds.

Yes, but for those who do outperform, the outperformance can be largely explained by factors!

To evaluate factors in terms of how individual investors might benefit from them, I wouldn't compare actively-managed funds to passively-managed funds, but a passively-managed fund that loads on one or more factors besides market beta to one that loads only on market beta (e.g., a total market cap-weighted fund). I showed an example of that above.

"The beauty of owning the market is that you eliminate individual stock risk, you eliminate market sector risk, and you eliminate manager risk. -- There may be better investment strategies than owning just three broad-based index funds but the number of strategies that are worse is infinite." -- Jack Bogle

Can't argue with this!

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

Theoretical
Posts: 1439
Joined: Tue Aug 19, 2014 10:09 pm

Re: "Factor investing"

Post by Theoretical » Fri Sep 09, 2016 1:52 pm

Kevin M wrote: I am skeptical too, but for the same reason I'm skeptical that stocks (factor = beta) will outperform bonds (factors = term and credit) over any particular period of interest to me, no matter how long.

Kevin
To me, that's been lynchpin argument for me to consider other factors, especially because Beta can also and has go(ne) negative or underperform(ed) for long stretches.

jay22
Posts: 705
Joined: Thu Aug 23, 2012 8:56 am
Location: Sacramento, CA

Re: "Factor investing"

Post by jay22 » Fri Sep 09, 2016 3:07 pm

larryswedroe wrote:
I would add that if you don't invest in factors other than market beta than you have all your eggs in that one risk basket, a basket which has the risk of performing very poorly for a very long time. Personally I would rather have a portfolio that is diversified across different sources of risk/return, where there is strong evidence that one should expect an ex-ante premium.
Larry, that is an interesting thought. So you're saying everyone should have funds which correspond to all other factors in their portfolio?

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Fri Sep 09, 2016 3:09 pm

Theoretical
EXACTLY, while beta has high odds of success for whatever the time frame, there are still long periods of underperformance. Even at 20 years in US it's 4%. Note it's historical odds of underperformance at 1, 3, 5 and 10 years are worse than for value, and just slightly better at 20 years by 2%. I would add profitability has higher odds of outperformance than beta at all horizons historically, and quality also with exception of 20 years where tied.
Finally a portfolio of factors has lower odds of underperformance at all horizons than even the best factor at any horizon, and the Sharpe ratios are dramatically higher for portfolio of factors than for beta, not even in same ballpark.
The idea IMO is to diversify sources of risk in world where we don't have clear crystal balls. The more confident you are in a factor likely having a premium the more weight one might put on it, and vice versa. So I have more confidence in value and beta than profitability/quality/momentum so I weight them more. But I still want some exposure to the others as principle of diversification.

It's why I also add other sources of risk, including a new fund that gains exposure to the reinsurance business. It's not a mutual fund though it's a 40 Act fund, with daily pricing but only quarterly liquidity (hence it's called an interval fund). Should have equity like returns as a business (same business as Warren Buffett is in with his reinsurance company) and obviously no correlation to stocks or bonds. Just another different source of risk with logical/intuitive premium

Hope that is helpful
Larry

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Fri Sep 09, 2016 3:16 pm

Jay
NO, I would not say that though I believe everyone should CONSIDER that. The reason is tracking error regret, a dread disease which causes investors to abandon their plans. ALL factors or asset classes go through long periods of underperformance. To "prove" my point here's example IMO which is the poster boy for the disease which occurs when a diversified portfolio underperforms a popular retail benchmark such as the S&P 500.

Unfortunately, it’s my experience that when contemplating investment returns, the typical investor considers three years a long time, five years a very long time, and 10 years an eternity. However, financial economists know that when it comes to the returns of risky assets periods as short as three or five years, or even 10, should be considered nothing more than noise. No more proof is required than over the nine years ending in 2008, the S&P 500 underperformed riskless one-month Treasuries by a compound 6.7% a year, a cumulative underperformance of 59%. And they underperformed five-year Treasuries by 10.8% a year, a cumulative underperformance of 115%. Investors in stocks should not have lost faith in their belief that stocks should outperform safer Treasuries due to that experience. And most didn’t. Yet, when small stocks or value stocks underperform for a few years, As Value stocks did from 2008 through 2015 when they underperformed growth stocks by just 2% a year, or a cumulative 16%, they begin to question their diversification strategy and lose discipline. You see this right here on Bogleheads despite the overwhelming evidence of value premiums all around the world for almost 100 years of data in US and about 50 worldwide, and across industries, sectors and even asset classes (whatever is cheap outperformance whatever is expensive). And there are excellent explanations for the value premium both risk based and behavioral which should give confidence in why it should persist.

The bottom line then is that if you don't have a VERY STRONG BELIEF SYSTEM in why the premium should be EXPECTED to outperform (ex-ante) then IMO you should not invest because the virtual certainty is that it will go for long time where it underperforms and you will catch the disease of tracking error regret and abandon your plan, usually just at the point when the premium returns.

Hope that helps
Larry

User avatar
aj76er
Posts: 603
Joined: Tue Dec 01, 2015 11:34 pm
Location: Portland, OR

Re: "Factor investing"

Post by aj76er » Fri Sep 09, 2016 3:26 pm

nisiprius wrote: Persistence: extra risk-adjusted return emerges from the interplay of characteristics and relationships between the asset classes that may not be any more persistent than return itself. Bill Miller beat the S&P 500 in return for fifteen years in a row, but that return did not persist. "Commodities" (using whatever dataset PortfolioVisualizer uses) has a U.S. market correlation from 1972 to 2006 of -0.17, and improved the risk-adjusted-return hypothetical backtested portfolios enough to convince fund companies to start including them in target-date funds, but that correlation did not persist; for 2007 through 2015, it was 0.62.

Rarity: The second is that it is not enough for two assets to have low correlation. In order to see a meaningful benefit, they have to have a) genuinely low correlation, not just "less than 1," and the big magic occurs only when you get to negative correlation; b) roughly comparable volatility; cash has zero correlation with stocks, but does not improve (nor harm) the risk-adjusted return of stocks because it has zero volatility; c) most important, and most often lacking, both assets need to have a decent and roughly comparable return.
But even if a -> c are not completely met, an asset with correlation <1 would provide some higher risk-adjusted return to a portfolio as a whole if 1) periodic rebalancing is done; and 2) reversion to the mean occurs at some point over the portfolio's holding period.

Thus, while commodities no longer have a -0.17 correlation, a 0.62 correlation still provides better long-term risk-adjusted returns, does it not? It may take a longer holding period, and at least one reversion to the mean, but assuming those occur isn't it still beneficial?

And thank you for such a thoughtful post, as always :).
"Buy-and-hold, long-term, all-market-index strategies, implemented at rock-bottom cost, are the surest of all routes to the accumulation of wealth" - John C. Bogle

lack_ey
Posts: 6635
Joined: Wed Nov 19, 2014 11:55 pm

Re: "Factor investing"

Post by lack_ey » Fri Sep 09, 2016 4:16 pm

aj76er wrote:Thus, while commodities no longer have a -0.17 correlation, a 0.62 correlation still provides better long-term risk-adjusted returns, does it not? It may take a longer holding period, and at least one reversion to the mean, but assuming those occur isn't it still beneficial?
No, it depends on risk/return for the assets. If you're looking to add an asset X to portfolio A and X has worse risk/return than A, then for X to help, X needs to have a high enough risk/return relative to the risk/return of A and its correlation with A. The higher the correlation with A, the higher X's risk/return must be to be worthwhile. Assuming normal distributions, I think it was that the ratio of the Sharpe ratios of X and A has to be greater than the correlation between X and A, if you want to improve the combination's Sharpe ratio over A.

e.g. when A has Sharpe of 0.4, X and A have correlation of 0.6, then X needs a Sharpe ratio of 0.24 or better. Assuming it does, as you add more X to the asset allocation, your new allocation A' is a mix of A and X with greater correlation to X and higher Sharpe than A, so the benefit of adding even more X keeps dropping off and eventually turns negative. Anyway, with non-normal distributions it's a bit more complicated but I think this is good enough for some minor intuition.

If you are looking for higher return for a given level of risk or lower risk for a given level of return, that is even more difficult than the above. After all, with the above formulation if you had an asset with zero correlation but very low Sharpe, adding some to your portfolio would reduce risk and return (but risk more than return). But simply shifting most asset allocations towards bonds (all the ones that aren't very heavy on bonds already) would also increase Sharpe while reducing both risk and return.

With commodities you also have to wonder about the long-term rate of return. Particularly when accessed as everyone pretty much has to through futures, if contango is relatively persistent on average then the rate of return might well not even beat the risk-free rate. You could have negative Sharpe potentially.

User avatar
aj76er
Posts: 603
Joined: Tue Dec 01, 2015 11:34 pm
Location: Portland, OR

Re: "Factor investing"

Post by aj76er » Fri Sep 09, 2016 4:28 pm

lack_ey wrote:
aj76er wrote:Thus, while commodities no longer have a -0.17 correlation, a 0.62 correlation still provides better long-term risk-adjusted returns, does it not? It may take a longer holding period, and at least one reversion to the mean, but assuming those occur isn't it still beneficial?
No, it depends on risk/return for the assets. If you're looking to add an asset X to portfolio A and X has worse risk/return than A, then for X to help, X needs to have a high enough risk/return relative to the risk/return of A and its correlation with A. The higher the correlation with A, the higher X's risk/return must be to be worthwhile. Assuming normal distributions, I think it was that the ratio of the Sharpe ratios of X and A has to be greater than the correlation between X and A, if you want to improve the combination's Sharpe ratio over A.

e.g. when A has Sharpe of 0.4, X and A have correlation of 0.6, then X needs a Sharpe ratio of 0.24 or better. Assuming it does, as you add more X to the asset allocation, your new allocation A' is a mix of A and X with greater correlation to X and higher Sharpe than A, so the benefit of adding even more X keeps dropping off and eventually turns negative. Anyway, with non-normal distributions it's a bit more complicated but I think this is good enough for some minor intuition.

If you are looking for higher return for a given level of risk or lower risk for a given level of return, that is even more difficult than the above. After all, with the above formulation if you had an asset with zero correlation but very low Sharpe, adding some to your portfolio would reduce risk and return (but risk more than return). But simply shifting most asset allocations towards bonds (all the ones that aren't very heavy on bonds already) would also increase Sharpe while reducing both risk and return.

With commodities you also have to wonder about the long-term rate of return. Particularly when accessed as everyone pretty much has to through futures, if contango is relatively persistent on average then the rate of return might well not even beat the risk-free rate. You could have negative Sharpe potentially.
Thank you! That explains it very well.
"Buy-and-hold, long-term, all-market-index strategies, implemented at rock-bottom cost, are the surest of all routes to the accumulation of wealth" - John C. Bogle

User avatar
nedsaid
Posts: 10644
Joined: Fri Nov 23, 2012 12:33 pm

Re: "Factor investing"

Post by nedsaid » Fri Sep 09, 2016 4:31 pm

If my portfolio outperforms its blended benchmark it is all due to my good looks, talent, smarts, and factors. If my portfolio underperforms, it is just bad luck. :D
A fool and his money are good for business.

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Fri Sep 09, 2016 4:37 pm

Few thoughts to add to Lackey
Contango is only an issue for hard to store commodities. Easily stored ones are pure arbitrages between interest rates and storage costs--like gold. So you don't have that issue. That is the case for most of the commodities, but not energy, which is why I would avoid futures where energy is the largest percentage of the portfolio, like GSCI. This is not the case with the others.

With that said, if you have persistent large contango you should expect negative real returns. Over time the energy complex had averaged about 0, some periods of backwardation and some contango. That has changed to mostly large contango for most of the last 8 years now. It could be "financialization" or something else (like inventory levels). But it seems hard to me to say it's purely financialization when I think it was in 2014 or maybe it was 15 when it swung from big contango to big backwardation and financialization had not changed. At any rate that has accounted for much of the poor performance of CCF, especially for energy heavy strategies. To me that should not logically continue (just as long periods of large backwardation should not) because losses would lead to outflows, reducing the contango. But my crystal ball is cloudy

Again, with commodities you have to understand that correlations are time/regime varying. They are negative with bonds, but positive with stocks when you have a negative demand shock (like in recession) and negative with stocks when have negative supply shock. So it depends on the regime. Also deflation generally bad for both and inflation generally good for commodities, but not so for stocks.

Hope that helps
Larry

User avatar
nisiprius
Advisory Board
Posts: 37048
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: "Factor investing"

Post by nisiprius » Fri Sep 09, 2016 5:57 pm

lack_ey wrote:...Assuming normal distributions, I think it was that the ratio of the Sharpe ratios of X and A has to be greater than the correlation between X and A, if you want to improve the combination's Sharpe ratio over A...
Yes. For example, if the Sharpe ratio of A is zero, then the correlation has to be negative.

However, as with (as with other aspects of MPT) does not depend on the distributions being normal. I derived this for myself using high-school algebra--it wasn't easy for me and I struggled for quite a while but it actually was high-school algebra... with some equations that were long enough that I was constantly making mistakes in simplifying them. But normality didn't enter into it anywhere.

I'd really like to know: did you learn this somewhere else, and if so, where (thus confirming that I did it correctly), or are you referencing my own postings here? I'm really pretty sure it's correct "experimentally," cross-checking with a different home-brew tool, but...
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.

lack_ey
Posts: 6635
Joined: Wed Nov 19, 2014 11:55 pm

Re: "Factor investing"

Post by lack_ey » Fri Sep 09, 2016 6:30 pm

nisiprius wrote:
lack_ey wrote:...Assuming normal distributions, I think it was that the ratio of the Sharpe ratios of X and A has to be greater than the correlation between X and A, if you want to improve the combination's Sharpe ratio over A...
Yes. For example, if the Sharpe ratio of A is zero, then the correlation has to be negative.

However, as with (as with other aspects of MPT) does not depend on the distributions being normal. I derived this for myself using high-school algebra--it wasn't easy for me and I struggled for quite a while but it actually was high-school algebra... with some equations that were long enough that I was constantly making mistakes in simplifying them. But normality didn't enter into it anywhere.

I'd really like to know: did you learn this somewhere else, and if so, where (thus confirming that I did it correctly), or are you referencing my own postings here? I'm really pretty sure it's correct "experimentally," cross-checking with a different home-brew tool, but...
I picked it up from your post here, but I did do some spot checks with a few examples using a spreadsheet I already set up to do some MPT portfolio calculations.

I didn't do the derivation for myself, though.

Call_Me_Op
Posts: 7063
Joined: Mon Sep 07, 2009 2:57 pm
Location: Milky Way

Re: "Factor investing"

Post by Call_Me_Op » Sat Sep 10, 2016 8:21 am

I think one thing that Nisi missed in his fine post above is that asset class returns are not Gaussian. That is, you can have higher return with higher risk but have similar (or at least proportionally-smaller) maximum draw-down. This, I believe, is the key appeal to factor investing. You can place a smaller proportion of your assets into stocks and achieve similar expected return with far smaller maximum draw-down.
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein

User avatar
nisiprius
Advisory Board
Posts: 37048
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: "Factor investing"

Post by nisiprius » Sat Sep 10, 2016 8:52 am

lack_ey wrote:...I picked it up from your post here, but I did do some spot checks with a few examples using a spreadsheet I already set up to do some MPT portfolio calculations...
Thank you very much, I really appreciate that. I should probably mention that I got the idea myself from a poster some years ago who said he'd done the algebra for the interesting special case where one asset has zero Sharpe ratio (and thus only help the portfolio if the correlation is negative). Wait, now that I actually have searched for it I find that he did a lot more... I was unconsciously repeating his work. It is by gw, here.
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.

paul merriman
Posts: 36
Joined: Fri Oct 25, 2013 1:00 am

Re: "Factor investing"

Post by paul merriman » Sat Sep 10, 2016 9:42 am

I don't consider "factor" investing an attempt to beat the market. I consider the addition of other than large cap blend (S&P or Total Stock Market) to simply be different markets. I'm not worried about beating any market but I am attempting to earn the best return I can within my risk tolerance. If I were in the accumulation stage of life, I would expect dollar cost averaging into a group of productive asset classes as one way to do better than the average investor but not better than the market. I sponsor Personal Investing 216 at Western Washington University. These students get almost 40 hours of education about how to make a lifetime of good investment decisions. By the way, the class does not discuss spend a minute on how to select a god stock but many hours about how to select a productive asset class. Those students are very likely to have at least 25% of their portfolio dedicated to index funds or ETFs that represent small cap value. Here is one of the many tables they get in the class
http://paulmerriman.com/four-fund-solution-table/. If the future looks anything like the past stocks will likely make more than bonds, small stocks will likely make more than large, value stocks will make more than growth and a well diversified portfolio can have significant holdings in REITs, emerging markets and other international asset classes. Since I don't trust the future to look like the past, I have approximately equal positions in all these equity asset classes.

I like the comment "that's good enough for me," referring to accepting whatever the outcome of large cap blend will make. I have no problem with that stance. At 72 I have half my investments in bonds and half in equities and that's "good enough" for me. I take out 5% of my portfolio the first of each year and that's "good enough" for my wife and me to enjoy our retirement. I pay an advisor to oversee my portfolio and take care of all of the details that I don't have an interest in monitoring. I might make more or less without an advisor but having one is "good enough" for us. I know at least 10 things I could do to make more and leave more but what we are doing is "good enough." For those who follow my marketwatch.com articles I will write a future articles on those 10 ways I could make more.

Since I retired almost 5 years ago I have helped a lot of people find a place for their money where it could be "good enough." I suggest many of those people put their long term investments in a combination of Vanguard Wellington and Wellesley. I have never had one of them come back and ask for a better recommendation. For those really interested in how I build a portfolio, I will have a free 2 1/2 hour video on the topic in a couple of weeks.

Random Walker
Posts: 3298
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Factor investing"

Post by Random Walker » Sat Sep 10, 2016 11:26 am

Paul,
There is a tremendous amount of wisdom in your "good enough". Reminds me of a story I read in one of my investing books, probably Bernstein or Swedroe, where a guy who had retired to Boca Raton from the cold north east was being interviewed. He was asked how his investments had done compared to the market. His answer was "who cares! I made it to Boca!" :-)

Dave

pkcrafter
Posts: 13126
Joined: Sun Mar 04, 2007 12:19 pm
Location: CA
Contact:

Re: "Factor investing"

Post by pkcrafter » Sat Sep 10, 2016 12:12 pm

Paul (Merriman), thanks for a very informative and useful post. Good enough is a real key to investing success as it dismisses useless attempts to try for perfection, which cannot be achieved because every decision involves a compromise of some sort.
By the way, the class does not discuss spend a minute on how to select a god stock but many hours about how to select a productive asset class.
This sounds interesting, are there any notes or videos available?

And finally, thanks for the video that will be available in a few weeks.


another Paul
When times are good, investors tend to forget about risk and focus on opportunity. When times are bad, investors tend to forget about opportunity and focus on risk.

User avatar
FIREchief
Posts: 2741
Joined: Fri Aug 19, 2016 6:40 pm

Re: "Factor investing"

Post by FIREchief » Sat Sep 10, 2016 12:30 pm

paul merriman wrote:

I take out 5% of my portfolio the first of each year and that's "good enough" for my wife and me to enjoy our retirement.

I pay an advisor to oversee my portfolio and take care of all of the details that I don't have an interest in monitoring.
Paul - hypothetically, if you pay that advisor 1% AUM, you are withdrawing 6% every year from a 50/50 AA. Consensus opinion on the forum is that with current circumstances, that is much more than a safe/sustainable withdrawal rate. Some have cited recent research suggesting the SWR is now less than 4%, more along the lines of 3%. Is your thinking that you are currently withdrawing much more than your "needs" and will greatly reduce your withdrawals in future years?
I am not a lawyer, accountant or financial advisor. Any advice or suggestions that I may provide shall be considered for entertainment purposes only.

jalbert
Posts: 3906
Joined: Fri Apr 10, 2015 12:29 am

Re: "Factor investing"

Post by jalbert » Sat Sep 10, 2016 12:42 pm

In response to the OPs link to 2003 about the efficient frontier... The efficient frontier 'works' (even after 2002) if implemented and systematically updated. Nearly all investors who use efficient frontier don't use it in that manner. They simply pick a buy and hold portfolio based on what the efficient frontier is currently. As the efficient frontier changes an investor needs to radically change their portfolio regularly which requires high (er) turnover. Investors who would monthly or quarterly update their portfolio to the efficient frontier would get the intended risk adjusted returns.
The true efficient frontier doesn't ever change. What gets computed is an estimated efficient frontier based on sample correlations from some historical time period. Because most historical financial data is highly biased by time period, these estimates are highly biased by time period, and it is these estimates that change significantly over time.

The problem is that you have no basis for selecting the historical time period that will provide best estimate of mean and variance of return in any particular projected future time period of interest, so there is no method of choosing which historical sample is relevant as an estimate of efficient frontier.

And even if we had a way of calculating the true efficient frontier (which would require knowing the actual probability distributions of future returns) all it tells us is that over a sufficiently long investment horizon, risk-adjusted return will be arbitrarily well approximated by expected risk-adjusted return. But it still could take 500 years.

Much risk is future, unpredictable events that never show up in historical samples, regardless of time period, and in truth, we don't know the probability distributions of future financial events or asset returns.
Last edited by jalbert on Sat Sep 10, 2016 2:20 pm, edited 1 time in total.

jalbert
Posts: 3906
Joined: Fri Apr 10, 2015 12:29 am

Re: "Factor investing"

Post by jalbert » Sat Sep 10, 2016 1:12 pm

One key thought here re past performance isn't predictive. That refers to the performance of ACTIVE MANAGERS, not factors.
There is no force of nature that requires past performance of any asset class or strategy to be predictive, including factor returns.

We've already seen that the small-cap factor was a liquidity risk premium that has mostly dissipated in risk-adjusted terms as small cap stocks became more liquid. And the supposedly superior risk-adjusted return of small caps in the past did not take liquidity risk into account (I don't think Fama & French considered trading costs or bid-ask spreads) so the true risk-adjusted return of small caps may never have actually exceeded that of the broad market.

If you want to bet that tilting to value or small-cap value will enhance risk-adjusted return, that's a reasonable thing to do. You may well be rewarded for doing so. But none of the research claiming superior risk-adjusted returns of such a strategy is solid enough for such a claim to be well established. Small-cap stocks were far too illiquid in the 1930's to use that time frame to try to understand how small-cap value stocks would behave today if we had serious deflation moving forward, an example of an unquantifiable risk that limits our understanding of expected future risk-adjusted returns.
Last edited by jalbert on Sat Sep 10, 2016 2:56 pm, edited 1 time in total.

Random Walker
Posts: 3298
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Factor investing"

Post by Random Walker » Sat Sep 10, 2016 2:48 pm

No one will ever know where the efficient frontier is. It will always depend on past time period chosen and we invest looking forward. But we can increase the likelihood that we are close to it by multi asset class investing / diversifying across sources of return.

Dave

User avatar
Taylor Larimore
Advisory Board
Posts: 27606
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: "Factor investing"

Post by Taylor Larimore » Sat Sep 10, 2016 4:08 pm

Random Walker wrote:No one will ever know where the efficient frontier is. It will always depend on past time period chosen and we invest looking forward. But we can increase the likelihood that we are close to it by multi asset class investing / diversifying across sources of return.

Dave
Dave:

Total market index funds include nearly ALL asset-classes.

Mr. Bogle likes to say: "Don't look for the needle. Buy the haystack."

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

larryswedroe
Posts: 15711
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: "Factor investing"

Post by larryswedroe » Sat Sep 10, 2016 4:33 pm

Taylor, they do include all the asset classes, but only have exposure to one factor, market beta. That's IMO an important distinction since we know that market beta only explains about 2/3 of the variation of returns of diversified portfolios. The other few factors that matter explain almost all the rest and all have low to negative correlation with market beta, providing a diversification benefit, as well as providing a premium. Note the premium can be used to lower beta exposure and add more TERM risk, another diversifying factor and that creates not only more of a diversified portfolio by factors but lowers the SD of the portfolio historically and produces higher Sharpe ratio.
So the way to think about it is why buy ONE haystack, beta, when you can diversify and buy many haystacks, each with premiums and low correlations.
Best wishes
Larry
Last edited by larryswedroe on Sat Sep 10, 2016 9:49 pm, edited 1 time in total.

Random Walker
Posts: 3298
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Factor investing"

Post by Random Walker » Sat Sep 10, 2016 5:28 pm

I think the reference to "buy many haystacks rather than just one haystack" really summarizes the distinction between the TSMers and the Tilters.

Dave

User avatar
FIREchief
Posts: 2741
Joined: Fri Aug 19, 2016 6:40 pm

Re: "Factor investing"

Post by FIREchief » Sat Sep 10, 2016 7:19 pm

larryswedroe wrote:Taylor, the (sic) do include all the asset classes, but only have exposure to one factor, market beta. That's IMO an important distinction since we know that market beta only explains about 2/3 of the variation of returns of diversified portfolios. The other few factors that matter explain almost all the rest and all have low to negative correlation with market beta, providing a diversification benefit, as well as providing a premium. Note the premium can be used to lower beta exposure and add more TERM risk, another diversifying factor and that creates not only more of a diversified portfolio by factors but lowers the SD of the portfolio historically and produces higher Sharpe ratio.
So the way to think about it is why buy ONE haystack, beta, when you can diversify and buy many haystacks, each with premiums and low correlations.
Best wishes
Larry
Well, it is good to finally hear the "counterpoint." That said, I'm with Taylor on this one. :sharebeer
I am not a lawyer, accountant or financial advisor. Any advice or suggestions that I may provide shall be considered for entertainment purposes only.

User avatar
JoMoney
Posts: 6257
Joined: Tue Jul 23, 2013 5:31 am

Re: "Factor investing"

Post by JoMoney » Sat Sep 10, 2016 7:32 pm

Random Walker wrote:I think the reference to "buy many haystacks rather than just one haystack" really summarizes the distinction between the TSMers and the Tilters.

Dave
I think that's the 'story' tilters want to use to justify the data mining and performance chasing. The reality is, people are still buying the same stocks in the same mutual funds that everybody else is, although maybe using different "stories" and theories to justify the way they choose to slice it up. These stocks are all very highly correlated, and offer no real surety in their "diversification" the way cash or a bond would.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

Random Walker
Posts: 3298
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Factor investing"

Post by Random Walker » Sat Sep 10, 2016 7:58 pm

I tilt, and I think it's fair to say that I don't depend on data mining or performance chasing. One shouldn't tilt if that is their only rationale. To tilt with conviction, one needs to look forward with a steadfast belief in a risk or persistent behavioral story. The data though is quite powerful too though.

Dave

lack_ey
Posts: 6635
Joined: Wed Nov 19, 2014 11:55 pm

Re: "Factor investing"

Post by lack_ey » Sat Sep 10, 2016 8:24 pm

Random Walker wrote:I think the reference to "buy many haystacks rather than just one haystack" really summarizes the distinction between the TSMers and the Tilters.

Dave
It's more like grabbing sections of a haystack rather than the whole thing, with the idea that different parts have different characteristics (which is fairly indisputable). The argument is in whether these different characteristics are better in any real way, potentially useful in portfolio construction.

JoMoney wrote:I think that's the 'story' tilters want to use to justify the data mining and performance chasing. The reality is, people are still buying the same stocks in the same mutual funds that everybody else is, although maybe using different "stories" and theories to justify the way they choose to slice it up. These stocks are all very highly correlated, and offer no real surety in their "diversification" the way cash or a bond would.
You want the story that is closer to the truth, whatever that is. All investing theory is based on understandings of past data, encompassing economic and statistical analyses.

If your standard for acting is surety or near surety, then I think it is much too high. In general, it can make sense to play the odds with educated guesses. You're not going to be right on every count but there may be some edge available. What we know is that, for example, chasing hot active managers has not been a successful strategy and is probably a bad idea for the future; other strategies have a higher chance of success and some ex-ante above 50%. Maybe. Probably?


Backing up a lot, even in a CAPM stock world, you can justify tilting stocks rather than owning the whole market. If your desired risk level about corresponds to 60% stocks (this part with a beta of 1), 40% bonds, it might be more efficient to instead use a broadly diversified slice of stocks with a beta of say 1.2 with almost zero additional idiosyncratic risk over the total stock market. You'd use 50% of the high-beta stocks and 50% bonds, or maybe something more like 49/51 to reach a risk level equivalent with the plain 60/40. But it would probably be better overall, with a little more even contribution of risks and better risk/return than the plain 60/40.

A more modern factor-centric analysis of the above would suggest it's a bad idea because high beta has been generally not a good part of the market, with lower risk/return than expected. The opposite of low beta and kind of opposite to low volatility. So you might use non-market factors to increase risk and expected return on the stock side rather than load up on high beta stocks.

User avatar
investorguy1
Posts: 543
Joined: Mon Nov 24, 2014 7:13 pm

Re: "Factor investing"

Post by investorguy1 » Sat Sep 10, 2016 8:29 pm

larryswedroe wrote:Taylor, the do include all the asset classes, but only have exposure to one factor, market beta.
1. Larry, I don't understand why a market weight portfolio only has exposure to one factor (market beta). It seems to me that market beta is really made up of all the factors. When you hold the market you are holding all possible factors in proportion to the market average. The bundle of all those factors in proportion to the market is market beta.

So for an average person holding the market should be fine. I could understand someone not average wanting different exposure for example someone who runs a small cap US mutual fund may want to reduce their exposure to that asset class.

2. I know that factors can persist after becoming well known. However their popularity is increasing at an extremely fast rate. Virtually every mutual fund and ETF company has or is coming out with smart beta funds (Vanguard, iShares, Fidelity, Goldman, powershares, investco, wisdom tree, Oppenheimer, legg mason, t. rowe just to name a few). Low volatility funds have exploded in size. So at some point couldn't these strategies get over valued leading to lower returns in the future?

3. I haven't done a scientific study but looking at many of these smart beta funds track records they aren't looking too good. I would be very surprised if any where near half have beaten low cost index funds in similar categories.

4. There is something very appealing, maybe even elegant about holding the market. You are guaranteed to beat the majority of investors (no other strategy offers this), costs and taxes are minimized and so is time, effort and tracking error regret.

lack_ey
Posts: 6635
Joined: Wed Nov 19, 2014 11:55 pm

Re: "Factor investing"

Post by lack_ey » Sat Sep 10, 2016 8:48 pm

investorguy1 wrote:
larryswedroe wrote:Taylor, the do include all the asset classes, but only have exposure to one factor, market beta.
1. Larry, I don't understand why a market weight portfolio only has exposure to one factor (market beta). It seems to me that market beta is really made up of all the factors. When you hold the market you are holding all possible factors in proportion to the market average. The bundle of all those factors in proportion to the market is market beta.

So for an average person holding the market should be fine. I could understand someone not average wanting different exposure for example someone who runs a small cap US mutual fund may want to reduce their exposure to that asset class.
I'm not Larry but I can never tell if people are asking this seriously or facetiously. I've seen some of both.

In case any is actually wondering, the other factors are set up and defined as performance differences between parts of the market. It is by construction that the total market doesn't have anything but the market factor. That doesn't mean the total market is deficient in any way. It means that it's not tilted.
investorguy1 wrote:2. I know that factors can persist after becoming well known. However their popularity is increasing at an extremely fast rate. Virtually every mutual fund and ETF company has or is coming out with smart beta funds (Vanguard, iShares, Fidelity, Goldman, powershares, investco, wisdom tree, Oppenheimer, legg mason, t. rowe just to name a few). Low volatility funds have exploded in size. So at some point couldn't these strategies get over valued leading to lower returns in the future?
What really matters is the net amount of money being invested this way and crowding up the trades, so to speak. The number of funds springing up is kind of an indirect measure. The effect would be evidenced by an increase in valuations in these parts of the market as their prices are bid up. This is perhaps what we see with some factors but not really for others. So while the concern is valid and some have brought it up (see: Rob Arnott), that doesn't particularly seem like it's actually happening to such a degree that we're all getting ready for the day of reckoning.

Also, note that the performance of a factor, even when bid up, can be decent, even during periods of times when the valuations are falling. The relationship between factor valuations and future returns varies by the factor and might not be as strong as you'd think, in part from factor constituent turnover.
investorguy1 wrote:3. I haven't done a scientific study but looking at many of these smart beta funds track records they aren't looking too good. I would be very surprised if any where near half have beaten low cost index funds in similar categories.
What do you count under the smart beta category? Do cap weighted factor tilts count? For the non-cap-weighted stuff, which some count as smart beta, some things like the RAFI-based funds did fairly well over the last 10 years, even over a relatively poor period for value. A lot of the category is pretty new yet. I don't know full statistics about this stuff, though.
investorguy1 wrote:4. There is something very appealing, maybe even elegant about holding the market. You are guaranteed to beat the majority of investors (no other strategy offers this), costs and taxes are minimized and so is time, effort and tracking error regret.
Elegance and tracking error regret are kind of behavioral considerations. Similarly, does it really matter what the majority of investors make and whether you beat them? It's relevant broadly but I think we've mostly been talking about raw risk and return considerations. Once you figure out what something does, then you can examine if you can stick with it and if it's worth the time, mental energy, and so on.

You should always consider costs and taxes, yes. An alternative strategy has to have benefits net of additional costs and taxes to be worth consideration. And that can't be evaluated unless you have some idea of the gross benefits to begin with, which you'd need to put numbers to.

Post Reply