"Minimize Fat-Tails" Portfolio Question

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

"Minimize Fat-Tails" Portfolio Question

Postby Call_Me_Op » Mon Apr 29, 2013 9:54 am

My understanding of a "Minimize Fat-Tails Portfolio" is a portfolio with a relatively small equity allocation that is highly tilted. The idea is to achieve returns comparable to more heavily equity-weighted portfolios but with less risk. With respect to the equity allocation, I assume we are not trying to be broadly-diversified (necessarily) but to take advantage of the "Fama-French factors" to maximize expected return. Is this the right way to think about it, or am I missing something?
Last edited by Call_Me_Op on Mon Apr 29, 2013 12:32 pm, edited 1 time in total.
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein
Call_Me_Op
 
Posts: 4266
Joined: 7 Sep 2009
Location: Milky Way

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Mon Apr 29, 2013 10:22 am

The strategy is based on simple idea, the more you tilt the less equity allocation you need to get the same return, that allows one to hold more SAFE (credit risk perspective) bonds, and you gain diversification of a different kind

If you own say a TSM and total international portfolio you are broadly diversified, owning the most stocks and have diversified across US, developed and EM and you own large and small and value and growth stocks. But, what most investors don't understand is that you have NO diversification at all across factors. The only exposure you have is to beta. Once you tilt you gain additional diversification to the factors of size and value.

So the Larry Portfolio (US SV, international SV and EM V) is diversified across the globe (US, developed and EM) and across all three factors with large exposure to size and value which have very low to no correlation to beta and each other. That provides a diversification benefit and in addition to the larger exposure to safe bonds cuts the tail risk. Historically it has been highly effective and has cut the left tail far more than it cut the right tail. We've been apply this type strategy for almost 20 years (with various degrees of tilting depending on the client).. Note I went to the extreme version, the Larry Portfolio, in 1998, bit early (:-))


Of course the more you tilt (allowing you to take less equity risk) the more of that dread disease known as tracking error regret. And old saying applies, if cannot stand the heat (of the TE) stay out of the kitchen. Not a strategy for those that confuse strategy and outcome--who don't understand that a strategy is right or wrong before you know the outcomes (at least in world with cloudy crystal balls).

Hope that is helpful

Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby Bradley » Mon Apr 29, 2013 11:02 am

larryswedroe wrote:If you own say a TSM and total international portfolio you are broadly diversified, owning the most stocks and have diversified across US, developed and EM and you own large and small and value and growth stocks. But, what most investors don't understand is that you have NO diversification at all across factors. The only exposure you have is to beta. Once you tilt you gain additional diversification to the factors of size and value.



Larry



Larry,
The factors of value and size ARE included in the total market portfolio and they help make up “market beta”. They are held in the same proportion as they exist “in total” of all marketable securities. Now if we follow your suggestion and add more (tilt/overweight) SCV we will pick up more size and value factor weighting, but to say you have NO diversification at all across factors is misleading and flat wrong.

Bradley
You can sum up any active fund manager’s presentation at an investor conference in one sentence: “We’re doing well, all things considered.”
Bradley
 
Posts: 486
Joined: 1 May 2007

Re: "Minimize Fat-Tails" Portfolio Question

Postby vesalius » Mon Apr 29, 2013 11:44 am

Bradley wrote:
larryswedroe wrote:If you own say a TSM and total international portfolio you are broadly diversified, owning the most stocks and have diversified across US, developed and EM and you own large and small and value and growth stocks. But, what most investors don't understand is that you have NO diversification at all across factors. The only exposure you have is to beta. Once you tilt you gain additional diversification to the factors of size and value.



Larry



Larry,
The factors of value and size ARE included in the total market portfolio and they help make up “market beta”. They are held in the same proportion as they exist “in total” of all marketable securities. Now if we follow your suggestion and add more (tilt/overweight) SCV we will pick up more size and value factor weighting, but to say you have NO diversification at all across factors is misleading and flat wrong.

Bradley

Bradley forgive me, but you seem to have become rather quixotic in regards to Mr. Swedroe. While I would have phrased it as either "very little" or "no practical" exposure to the size or value, given how they are dwarfed and overwhelmed in a total market fund, we are dancing on the head of a pin looking for reasons to disagree.
vesalius
 
Posts: 601
Joined: 13 Jul 2010
Location: Texas

Re: "Minimize Fat-Tails" Portfolio Question

Postby gt4715b » Mon Apr 29, 2013 11:57 am

Bradley wrote:Larry,
The factors of value and size ARE included in the total market portfolio and they help make up “market beta”. They are held in the same proportion as they exist “in total” of all marketable securities. Now if we follow your suggestion and add more (tilt/overweight) SCV we will pick up more size and value factor weighting, but to say you have NO diversification at all across factors is misleading and flat wrong.
Bradley


What Larry is saying is correct and only misleading if you don't understand the French Fama 3-Factor model. If you own TSM your factor loadings are 1,0,0 for Mkt, Small, and Value respectively. When we say Small and Value loading we are by definition talking about owning these in a higher/lower proportion than the market weighting.
gt4715b
 
Posts: 321
Joined: 11 Jun 2007

Re: "Minimize Fat-Tails" Portfolio Question

Postby Call_Me_Op » Mon Apr 29, 2013 12:12 pm

gt4715b wrote:
Bradley wrote:Larry,
The factors of value and size ARE included in the total market portfolio and they help make up “market beta”. They are held in the same proportion as they exist “in total” of all marketable securities. Now if we follow your suggestion and add more (tilt/overweight) SCV we will pick up more size and value factor weighting, but to say you have NO diversification at all across factors is misleading and flat wrong.
Bradley


What Larry is saying is correct and only misleading if you don't understand the French Fama 3-Factor model. If you own TSM your factor loadings are 1,0,0 for Mkt, Small, and Value respectively. When we say Small and Value loading we are by definition talking about owning these in a higher/lower proportion than the market weighting.


Clearly said, gt.
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein
Call_Me_Op
 
Posts: 4266
Joined: 7 Sep 2009
Location: Milky Way

Re: "Minimize Fat-Tails" Portfolio Question

Postby RNJ » Mon Apr 29, 2013 12:27 pm

Is it reasonable to look at the obverse, that is, the elimination or reduction of large and growth "factors"? With the TSM approach, one has exposure to all market factors (large, small, value and growth) which, taken together, equal beta. When one eliminates growth stocks and reduces large cap holdings one is left with a tilt away from these large and growth and towards value and small.

So, within a TSM approach, Bradley isn't wrong - the market holds all factors but in total market proportions. But we can't hold small and value in higher proportions - tilting - without eliminating or reducing large and growth. My understanding is that with tilting we are expressly not holding everything in equal proportions. We hold more of this and, as a result, less of that.

Re-staing the obvious?
RNJ
 
Posts: 517
Joined: 8 Apr 2013

Re: "Minimize Fat-Tails" Portfolio Question

Postby Call_Me_Op » Mon Apr 29, 2013 12:32 pm

RNJ wrote:Is it reasonable to look at the obverse, that is, the elimination or reduction of large and growth "factors"? With the TSM approach, one has exposure to all market factors (large, small, value and growth) which, taken together, equal beta. When one eliminates growth stocks and reduces large cap holdings one is left with a tilt away from these large and growth and towards value and small.

So, within a TSM approach, Bradley isn't wrong - the market holds all factors but in total market proportions. But we can't hold small and value in higher proportions - tilting - without eliminating or reducing large and growth. My understanding is that with tilting we are expressly not holding everything in equal proportions. We hold more of this and, as a result, less of that.

Re-staing the obvious?


RNJ,

I think reducing large and growth factors is the same as increasing small and value factors. The changes are measured with respect to market, and increasing proportion of one factor necessarily reduces the proportion of the other factors.
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein
Call_Me_Op
 
Posts: 4266
Joined: 7 Sep 2009
Location: Milky Way

Re: "Minimize Fat-Tails" Portfolio Question

Postby staythecourse » Mon Apr 29, 2013 12:51 pm

I must say we must be running out of stuff to argue about. EITHER WAY the result is the same TSM has factor loadings of 0 for small and value. One's who disagree should talk to Mr. French and Mr. Fama as they determined the criteria. Nuff said.

Now we can move on to the real arguments for or against tiling (once again :D ).

Good luck.
...we all think we're above average investors just like we all think we're above average dressers... -Jack Bogle
staythecourse
 
Posts: 2799
Joined: 3 Jan 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby rkhusky » Mon Apr 29, 2013 12:58 pm

The factors are variables in a set of equations, which are fit to past performance data.
rkhusky
 
Posts: 1346
Joined: 18 Aug 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby matjen » Mon Apr 29, 2013 1:07 pm

It is obvious that this portfolio is one of the more interesting ones out there and one that often brings up threads and questions. Moreover, Larry is a valued member of the community. I really think it should be in the wiki alongside 3 fund, core 4, etc. One thing I am struggling with (should be another thread) is how one moves to this portfolio effectively if they have won the game or nearly won it and have relatively large equity holdings (meaning large capital gains if sold) in taxable accounts. It is one thing to take a hit to reduce crazy ER expenses or to fix a horrid portfolio, but once you have Boglehead religion and have a reasonable and cheap portfolio...things get dicey it seems to me. Now you are taking a big tax hit to cut off the tail theoretically. Not as compelling an argument to me. Though next bear market my tune will change I am guessing. Just my .02
A man is rich in proportion to the number of things he can afford to let alone.
User avatar
matjen
 
Posts: 575
Joined: 21 Nov 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby hafius500 » Mon Apr 29, 2013 1:33 pm

vesalius wrote:Bradley forgive me, but you seem to have become rather quixotic in regards to Mr. Swedroe. While I would have phrased it as either "very little" or "no practical" exposure to the size or value, given how they are dwarfed and overwhelmed in a total market fund, we are dancing on the head of a pin looking for reasons to disagree.
(*)
gt4715b wrote:What Larry is saying is correct and only misleading if you don't understand the French Fama 3-Factor model. If you own TSM your factor loadings are 1,0,0 for Mkt, Small, and Value respectively. When we say Small and Value loading we are by definition talking about owning these in a higher/lower proportion than the market weighting.

Call_Me_Op wrote:Clearly said, gt

staythecourse wrote:I must say we must be running out of stuff to argue about. EITHER WAY the result is the same TSM has factor loadings of 0 for small and value. One's who disagree should talk to Mr. French and Mr. Fama as they determined the criteria. Nuff said.


Sorry, but it's you who does not understand factor models.

(*) For example, IIRC Russell classifes country markets ("TSM") as value or growth markets.
prior username: hafis50
hafius500
 
Posts: 155
Joined: 10 Sep 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby Clive » Mon Apr 29, 2013 2:23 pm

larryswedroe wrote:The strategy is based on simple idea, the more you tilt the less equity allocation you need to get the same return, that allows one to hold more SAFE (credit risk perspective) bonds

To expand upon what Larry said further, based on SImba's backtest spreadsheet data, if you multiply each assets annualised gain (CAGR) since 1972 by T-Bills Standard deviation and divide by the assets standard deviation and plot :

Image
(average 1.61)

If you consider volatility (standard deviation) to be a risk factor, then the above implies that when stock assets are levelled for risk, the rewards are broadly similar (lower end values of 1.27 for Small Cap Growth to upper end of 2.1 for REIT, Large Cap Value, Small Cap Value for the stock type assets).

Some stocks, perhaps small cap value for instance, might have higher debt to equity ratio's. If one stock has a debt/equity of 1 then its borrowed the same amount again as its equity value. Compare that to another stock that had a 2 debt/equity ratio and clearly the latter has more to invest and if the earnings exceed the cost of debt then the leverage was worthwhile and added value (gains).

If you have a more leveraged product, you can de-scale the amount held to achieve a similar result to a less leveraged product. ERX for instance is a 3x leveraged energy ETF, and if you hold one third of the amount you would have invested in the 1x energy ETF (IYE) then generally the outcome is much the same - but you're only holding 33% in 'stock' (ERX) instead of 100% (IYE)

Image
Image
Image

If you don't rebalance that investment back to target weightings periodically it will drift and you'll end up with tracking error. After a year for instance there might be 45% ERX and 55% TIP weightings, which wont track 100% IYE. Rebalancing periodically, perhaps once each year, will realign the two, so tracking error will be smaller. If you're only rebalancing once each year, then the most ERX can lose is the 33% invested amount, so your maximum loss (left tail risk) is lower than holding 100% in IYE (that could conceptually lose more than 33% over a year long period).

Zvi Bodie takes that further still, and rather than utilising moderate leverage (small cap value etc.) he suggest using less exposure to more leveraged products such as LEAPS (longer dated Options). Perhaps 90% TIPS, 10% LEAPS for instance.

You don't want to throw out diversification in the pursuit of leverage however. Diversification is generally a good thing, especially if one asset tends to zig as another zags.
Clive
 
Posts: 959
Joined: 13 Jun 2009

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Mon Apr 29, 2013 4:12 pm

The market BY DEFINITION has no loading on size and value. The reason is that small and value are LONG-SHORT portfolios.

The exposure to the small factor that TSM has because it owns small stocks is EXACTLY offset by the NEGATIVE exposure it has to the size factor via the large stocks it owns.

The same is true for value with the + exposure it has by owning value stocks is exactly offset by the negative exposure it has by owning growth stocks


Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby Bradley » Mon Apr 29, 2013 4:37 pm

The total market already incorporates average exposure to both the risks and the rewards of value and small cap investing. Investors who believe that the markets are efficient start by holding total-market portfolios which have average exposure to both the value and the small-cap style. The research shows there is a benefit to holding both small and value, domestic as well as international securities, in higher allocations than exist in a TM portfolio.........but, In one breath to say “you are broadly diversified, owning the most stocks and have diversified across US, developed and EM and you own large and small and value and growth stocks.” and then follow by saying, “But, what most investors don't understand is that you have NO diversification at all across factors.” is simply not an accuarate/acceptable conclusion. This forum is a great resource and misstatements left uncorrected or unaddressed can be accepted as truth by those who don't know any better. Many here understand what you were trying to say and appreciate your efforts.

Bradley
You can sum up any active fund manager’s presentation at an investor conference in one sentence: “We’re doing well, all things considered.”
Bradley
 
Posts: 486
Joined: 1 May 2007

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Mon Apr 29, 2013 5:02 pm

But, what most investors don't understand is that you have NO diversification at all across factors.” is simply not an accuarate/acceptable conclusion. This forum is a great resource and misstatements left uncorrected or unaddressed can be accepted as truth by those who don't know any better
.

Unfortunately there is no good way to address these type statements which are totally wrong.
So I'll just say that there is nothing inaccurate nor unacceptable about what I said. PERIOD.

Best wishes
Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby Bradley » Mon Apr 29, 2013 5:27 pm

Is it not the diversification across all factors that actually make up the Beta in the TSM?
You can sum up any active fund manager’s presentation at an investor conference in one sentence: “We’re doing well, all things considered.”
Bradley
 
Posts: 486
Joined: 1 May 2007

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Mon Apr 29, 2013 7:35 pm

s it not the diversification across all factors that actually make up the Beta in the TSM?


Read the definitions in my original posts. This is not a debatable points, just facts, definitions.

There is no diversification across factors in TSM. There is only exposure to BETA.

The mistake that many people make is that TSM is diversified across ASSET CLASSES, holding small and value stocks in their proportion, but has no exposure by FACTORS, again by definition TSM has none to size and value.

The factors have very low correlation to each other. BETA to size has been about 0.4. value to beta about 0.1 and size to value about 0. These are much lower correlations than the asset classes have to TSM.

Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby Bradley » Mon Apr 29, 2013 8:43 pm

Larry,

Thank you for your explanation and measured response.

Bradley
You can sum up any active fund manager’s presentation at an investor conference in one sentence: “We’re doing well, all things considered.”
Bradley
 
Posts: 486
Joined: 1 May 2007

Re: "Minimize Fat-Tails" Portfolio Question

Postby Beat The Street » Mon Apr 29, 2013 8:59 pm

Why doesn't the Larry Portfolio diversify further using a momentum loading?
5 out of every 4 people are bad at fractions.
Beat The Street
 
Posts: 165
Joined: 26 Sep 2012

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Mon Apr 29, 2013 10:18 pm

Beat the street

The Larry Portfolio eliminates the negative loading on MOM that value tends to have as the funds screen out negative MOM stocks (both the DFA and the Bridgeway funds I use do that)
As to MOM as separate fund in taxable account not convinced it adds further value beyond incorporating into a value strategy in a core type portfolio.

Now adding MOM to a high value tilt portfolio would likely reduce TE, but I don't care about that problem, though many do.

Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby Beat The Street » Mon Apr 29, 2013 10:32 pm

Sorry if I misunderstand this, but why should you just eliminate the negative momentum loading? Shouldn't you want a positive loading on momentum since the premium has been so large? Or no because it is believed to be an anomaly?
5 out of every 4 people are bad at fractions.
Beat The Street
 
Posts: 165
Joined: 26 Sep 2012

Re: "Minimize Fat-Tails" Portfolio Question

Postby steve roy » Mon Apr 29, 2013 11:03 pm

staythecourse wrote:I must say we must be running out of stuff to argue about. EITHER WAY the result is the same TSM has factor loadings of 0 for small and value. One's who disagree should talk to Mr. French and Mr. Fama as they determined the criteria. Nuff said.

Now we can move on to the real arguments for or against tiling (once again :D ).

Good luck.


Well, I'm in favor of the all-tile bathroom. I can only hope it won't be held against me.
User avatar
steve roy
 
Posts: 822
Joined: 13 May 2010

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Tue Apr 30, 2013 7:49 am

Beatthestreet
Strategies have no costs but implementing them does.
Eliminating negative MOM doesn't cost anything, it saves some costs by delaying trades you would otherwise make. Adding MOM would add costs in terms of trading costs/turnover and some loss of tax efficiency. Also adding MOM lowers the value exposure, no way around that. So you have to balance that.
And that's the question, would it add value. It might in large where transactions costs are small but perhaps not in small where they are large---thus just eliminating the negative MOM might be superior strategy.
And of course you have to consider the cost of the fund itself, and then rebalancing costs in taxable accounts---if you add a separate MOM fund
And then you have to decide how much you are willing to "bet" on something with no risk story, though a long history and persistence around the globe and across asset classes.

I hope that is helpful
Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby Beat The Street » Tue Apr 30, 2013 8:12 am

That is helpful, thank you sir.
5 out of every 4 people are bad at fractions.
Beat The Street
 
Posts: 165
Joined: 26 Sep 2012

Re: "Minimize Fat-Tails" Portfolio Question

Postby SquawkIdent » Tue Apr 30, 2013 3:38 pm

If you are using ETF's or index mutual funds to build this portfolio which ones would I look at?

I do not have assess to DFA Funds. Thanks.
User avatar
SquawkIdent
 
Posts: 179
Joined: 23 Dec 2008

Re: "Minimize Fat-Tails" Portfolio Question

Postby matjen » Tue Apr 30, 2013 6:49 pm

There is another thread going with some ideas. A few more in the past had some ideas as well.

viewtopic.php?f=1&t=115532&newpost=1682673
A man is rich in proportion to the number of things he can afford to let alone.
User avatar
matjen
 
Posts: 575
Joined: 21 Nov 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby Bongleur » Wed May 08, 2013 5:17 pm

Taleb has an article questioning the validity of our ability to measure. Comments from those who can follow the math?

Fat Tails and (Anti)Fragility
https://docs.google.com/file/d/0B_31K_M ... edit?pli=1
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
Bongleur
 
Posts: 2024
Joined: 3 Dec 2010

Re: "Minimize Fat-Tails" Portfolio Question

Postby LadyGeek » Wed May 08, 2013 7:54 pm

He really did publish that, here's his announcement on facebook: Nassim Nicholas Taleb | Facebook

I'm sure it's correct, but it seems to be overkill. IOW, interjection of higher-level math where a more fundamental approach would do. He states his point, shows the math, then moves on.

BTW: You should post this in one of the Taleb threads such as The anti-Taleb reviews Antifragile. Or, start a new thread. You'll get a much better response to your question.
To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
User avatar
LadyGeek
Site Admin
 
Posts: 15969
Joined: 20 Dec 2008
Location: Philadelphia

Re: "Minimize Fat-Tails" Portfolio Question

Postby Bongleur » Thu May 09, 2013 11:14 pm

That thread doesn't get into the question at all. The lecture I refer to seems to be showing that quite frequently the measures of risk used are actually inappropriate (which curve -- Guassian, Normal, etc) because we can't determine which is correct... so lots of our fundamental basis in thinking about investments is wrong. Especially the statistical methods appropriate to tail events.

Hoping Larry, King of Fat Tail Optimization, would opine.
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
Bongleur
 
Posts: 2024
Joined: 3 Dec 2010

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Fri May 10, 2013 10:39 am

Bongleur

Not sure what you are asking about, but hopefully this helps

We know that returns are not normally distributed. This isn't a new discovery by the way. In fact Fama's dissertation about 50 years ago was on this subject if my memory serves.

And because investors are on average fairly risk averse they care greatly, or should, about the left tail risk and thus investments that do poorly in bad times should care high premiums--and they do, in general.

With that in mind investors should be interested in portfolios that cut tail risk, especially if it imposes no costs in terms of expected returns, but even if it did they should logically be willing to pay some "insurance" premium to hedge that risk. (my argument for CCF for example and TIPS)

So when one constructs a portfolio one should consider these issues.

The "Larry Portfolio" does just that. And historically it has done it without paying a price in terms of lost expected returns and has cut the left tail risk far more than the right tail "risk" has been cut. That's because of how the risk factors work together, low correlation and when needed most high quality bonds tend to perform well. Finally, what helps is that the Larry Portfolio really takes a "risk parity" type approach, long before people began talking about risk parity and the strategy became popular.

The typical TSM 60/40 portfolio has the vast majority of its risks in one place--BETA, with very little in any other factor

Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby matjen » Fri May 10, 2013 10:47 am

larryswedroe wrote:
With that in mind investors should be interested in portfolios that cut tail risk, especially if it imposes no costs in terms of expected returns, but even if it did they should logically be willing to pay some "insurance" premium to hedge that risk. (my argument for CCF for example and TIPS)

So when one constructs a portfolio one should consider these issues.

The "Larry Portfolio" does just that. And historically it has done it without paying a price in terms of lost expected returns and has cut the left tail risk far more than the right tail "risk" has been cut. That's because of how the risk factors work together, low correlation and when needed most high quality bonds tend to perform well. Finally, what helps is that the Larry Portfolio really takes a "risk parity" type approach, long before people began talking about risk parity and the strategy became popular.


Larry I have been buzzing around your portfolio and theories for awhile. One issue I am not clear on is that we have had an incredible bond market for 30 years. Moving forward this almost certainly won't be the case (Note Nisi, I am not claiming a bond "crash!" :) ) With this assumption don't you think you would be more likely to cut more of the right tail moving forward?
A man is rich in proportion to the number of things he can afford to let alone.
User avatar
matjen
 
Posts: 575
Joined: 21 Nov 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby larryswedroe » Fri May 10, 2013 12:05 pm

Matjen
That's certainly a possibility. But with that said, investors should in theory be willing to give up more of the right tail to cut left tail risk because they are risk averse. And for any investor who has "already won the game", meaning they now have low marginal utility of wealth, the value of the right tail is very little compared to the cost of the left tail.
But I would also add this, the data I use to illustrate the value of the "Larry Portfolio" I show using 1 year Treasuries, so the portfolio did not really benefit from the big bond rally. Of course the results were much better (cut more of left tail and kept more of right tail if you held what I actually typically recommend, say a five year maturity, and absolute returns were quite a bit higher as well). But I show the data with 1 year for the reason you state, you should not expect the past to repeat here for bonds
Hope that is helpful
Larry
larryswedroe
 
Posts: 11202
Joined: 22 Feb 2007
Location: St Louis MO

Re: "Minimize Fat-Tails" Portfolio Question

Postby matjen » Fri May 10, 2013 12:56 pm

That is helpful Larry. Thank you for taking the time.
A man is rich in proportion to the number of things he can afford to let alone.
User avatar
matjen
 
Posts: 575
Joined: 21 Nov 2011

Re: "Minimize Fat-Tails" Portfolio Question

Postby grayfox » Fri May 10, 2013 6:47 pm

Bradley wrote:Is it not the diversification across all factors that actually make up the Beta in the TSM?

I don't think so. Beta is one factor, which explains something like 30% of a typical stocks returns, which is measured by R^2. (This is Sharpe's Rule of Thumb)
So a single-factor Beta model leaves a lot (~70%) unexplained for a typical stock, which shows up in either alpha term or the residual term.

Adding another factor, say size, should raise R^2, so that alpha or residuals get smaller.
Add another factor, say value, should raise R^2 further.
I have never seen what R^2 for a typical stock for the 3-factor model. My working assumption is that it's about 0.40 or 0.50

If you could kept adding more factors until R^2 = 1.00 then all the return of a stock would be explained by the model. But it's probably not possible to explain all the returns of a stock. There is always going to be some randomness that can't be explained by anything.

BTW: These factor models are not at all complicated and don't use any fancy math. For years I thought that they must be some fancy complex model until I read an article at thecalculatinginvestor that explained how to calculate everything. It turns out that they are just OLS regression models which most people should be familiar with from a basic college probability and statistics course. The betas are just the slopes and alpha is just the y-intercept. :oops:
Тише едешь, дальше будешь. (Quieter you-go, further you-will-be.)
User avatar
grayfox
 
Posts: 3831
Joined: 15 Sep 2007
Location: Anytown, USA

Re: "Minimize Fat-Tails" Portfolio Question

Postby LadyGeek » Fri May 10, 2013 7:10 pm

grayfox wrote:BTW: These factor models are not really that complicated and don't use any fancy math...

For those who want to deep-dive into the details (and see how the slope and Y-intercept is obtained), see the wiki: Fama-French three-factor model analysis

OLS = Ordinary least squares

For new investors, the portfolios and techniques described here deviate from the Boglehead's approach to invest in the total market, known as tilting. Be sure you understand the additional risks when considering this approach.
To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
User avatar
LadyGeek
Site Admin
 
Posts: 15969
Joined: 20 Dec 2008
Location: Philadelphia


Return to Investing - Theory, News & General

Who is online

Users browsing this forum: chicks, Clever_Username, HongKonger, JimB, JRA, Lauren Vignec, Lee Saage, lostcowboy, Louis Winthorpe III, nedsaid, nvboglehead, Rodc, Sard, SteveGo, stevewolfe and 74 guests