Matrix Redux: The Larry-Browne Permanent Portfolio

[grayfox]

In the voice of Donald la Fontaine: In a world where mundane lazy stock-and-bond portfolios rule, where buy-and-hold investors get crushed by falling markets, occasionally someone rises above the crowd...

Every investment book nowadays seems to recommend some combination of TSM, International, Value, Small Cap, REITs, intermediate bonds, etc., etc. Some lump all into TSM, others slice and dice. Couch potato, coffeehouse, Gone Fishin‘--there really isn’t a whole lot of difference among them: they are all mostly a combination of two broad asset classes stocks and bonds. Some lumped, some sliced. But not a one of those conventional portfolios provided much downside protection in 2008. They all took a shellacking. Their Kung Fu proved to be weak and not up to the task.

The Harry Browne Permanent Portfolio stands apart from these. It offers real diversification among four different asset classes: stock, gold, long-term treasuries, and cash. Harry Browne PP exhibited very powerful Kung Fu during the 2008 crash.

Now along comes the younger and quicker Larry with his portfolio. He allocates the bulk of his assets to safe Treasuries (maybe 70%), and the rest to very volatile and risky Small Cap Value and International Small Cap Value. Larry’s Portfolio also exhibited very powerful Kung Fu during 2008.

Harry Browne, the old-school martial artist, is like Agent Smith in the Matrix. He has defeated all who went up against him. Larry, like Neo in the Matrix, has shown that he can dodge bullets. Agent Smith thinks he is going to shut Neo down like he has done to so many others. In the most powerful scene, as a train bears down on Neo,

Agent Smith : You hear that, Mr Anderson? That is the sound of inevitability. The sound of your death. Goodbye, Mr Anderson.
Neo: My name is… NEO.

But Agent Smith has underestimated Neo’s powers. When Smith confronts Neo for one-on-one combat and shoots at him, Neo stops the bullets in mid flight using only his mind. Then Neo runs and dives into Smith and Smith is destroyed in a massive explosion of energy.

Or was he destroyed? It turns out he emerges as a more powerful hybrid of Smith and Neo. The most powerful Kung Fu ever unleashed.

That is what happened on the Bogleheads forum. The kung fu of Larry has been mixed with the jujitsu of Harry Browne. The result is The Larry-Browne Permanent Portfolio. It is like Harry Browne, except instead of allocating 25% to TSM, slice up the stock allocation 12.5% to SV and 12.5% to ISM.

The L-B PP has the same low volatilty has the original HB PP, but with a higher historical return. CAGR is 8.95 (HB) vs. 10.34 (LB)

Here are the return results presented by Trev H on another thread:

Here you go Roy...

Annual Returns year-by-year for Lumper PP vs Larry PP.

I put these returns into Excel and took the difference.

Code: Select all

         HB PP     L-B PP    LB-HB
YEAR     RETURN    RETURN    DIFF
1969               
1970      4.53      6.01     1.48
1971     14.04     18.21     4.17
1972     21.03     22.96     1.93
1973     16.02     15.17    -0.85
1974     14.21     14.72     0.51
1975      7.11     10.74     3.63
1976      7.97     12.36     4.39
1977      8.50     18.18     9.68
1978     15.47     21.02     5.55
1979     39.64     40.53     0.89
1980     11.23     11.74     0.51
1981     -4.85     -2.09     2.76
1982     18.12     18.92     0.80
1983      4.40      7.87     3.47
1984      2.71      2.79     0.08
1985     22.31     24.04     1.73
1986     23.99     21.33    -2.66
1987      9.04     11.15     2.11
1988      3.80      5.78     1.98
1989      9.22     10.07     0.85
1990      0.70     -0.40    -1.10
1991      8.27      7.83    -0.44
1992      1.21      2.43     1.22
1993     15.42     18.79     3.37
1994     -0.80      0.75     1.55
1995     12.90     10.50    -2.40
1996      3.08      2.60    -0.48
1997      3.06      0.49    -2.57
1998      9.20      4.00    -5.20
1999      5.55      2.86    -2.69
2000      1.67      6.73     5.06
2001     -1.31      3.77     5.08
2002      6.04      9.37     3.33
2003     14.65     18.79     4.14
2004      7.50     10.62     3.12
2005      9.48     10.54     1.06
2006     12.33     13.19     0.86
2007     13.87     11.46    -2.41
2008     -2.79     -2.10     0.69
2009      1.91      2.65     0.74

MIN      -4.85     -2.10    -5.20
MEAN      9.26     10.66     1.40
MAX      39.64     40.53     9.68
SD        8.49      8.69     2.80
NEG       4.00      3.00    10.00
POS      36.00     37.00    30.00

In forty years, HB had just four down years. L-B only three. The worst down year for HB was -4.85%, for L-B –2.10%. Standard deviation is about the same. L-B has a little higher average return than HB, 10.66% vs. 9.26%. When I plot the returns, the graphs overlay very closely, except for a few periods. But there is a slight bias upward for L-B compared to the old-school HB. L-B beat HB in 30 out of 40 years.

A new era has dawned in the world of investing.

Morpheus: You have to let it all go, Neo. All fear, doubt and disbelief. Free your mind.

[Roy]

In forty years, HB had just four down years. L-B only three. The worst down year for HB was -4.85%, for L-B –2.10%. Standard deviation is about the same. L-B has a little higher average return than HB, 10.66% vs. 9.26%.

If using S/D as the only measure of risk (it isn't) the added returns (1.4% more for only a bit more S/D) on the Larry variant are large. Backtesting is all we have on any of these models, but judged by the butcher's bill of that data, this is an impressive portfolio. For me there are two huge points:

It shows that holding just 25% stocks over this period (and using 75% in even just AAA bonds) accomplished the prime directive: don't lose. So the low beta is the main reason for this protection.

It also shows just how powerful the HB concept is because it uses true diversification with multiple volatile classes, any one of which can make huge movements across short periods (like last last year). And, as Craig has pointed out in the main thread, I think that is the central feature, and rebalancing among the 4 classes is paramount.

The Larry "tilt" offers an impressive risk/return improvement, but is secondary, in my view, to the othe two points.

Roy

[muck53]

Is not loosing, the new definition of winning?

I assume all your stats are based on calendar years? Perhaps I am wrong.

There is great appeal to me in even proposing that not loosing is the new definition of winning. I reasoned, back in the 1991 recession - that long term (over the next 20 years) I would be better off working at something, anything - then merely sitting home collecting unemployment and drinking beer, watching the flowers grow.
muck53

[stratton]

You sure this isn't an excuse/attempt to move the "betting" portion of investing money that Harry Browne said should be outside the PP into the portfolio?

I think you'd do better with four fund ultimate buy and hold portfolio of LB, SV, ILV, ISC instead of an all small cap bet. The returns are almost as good and using all three factors mean not losing big in any type of market. The problem is its four funds instead of one. Two if you do like Craigr with his 20% TSM and 5% Total Intl.

Paul

[Roy]

I think you'd do better with four fund ultimate buy and hold portfolio of LB, SV, ILV, ISC instead of an all small cap bet. The returns are almost as good and using all three factors mean not losing big in any type of market.

This would be my preference also with the HP PP.

Roy

[larryswedroe]

grayfox
I would not use Tbills (my assumption of what you used) but TIPS which should have significantly higher return and more of a diversification return as well

[hewhomustnotbenamed]

Voldemort says risk managing wizardry beats muggle kung-fu.

Precrash: 100% equity indexed annuity

Postcrash: 23.5% cash, 153% emerging market


commodity exposure coming soon

[grayfox]

I think you'd do better with four fund ultimate buy and hold portfolio of LB, SV, ILV, ISC instead of an all small cap bet. The returns are almost as good and using all three factors mean not losing big in any type of market. The problem is its four funds instead of one. Two if you do like Craigr with his 20% TSM and 5% Total Intl.

Paul

That sounds like a good approach. What would the Vanguard funds be?

6-1/4% LB VTSMX
6-1/4% SV VISVX
6-1/4% ILV VGTSX or VFWIX or VTRIX
6-1/4% ISC VINEX or VSS
25% GOLD GLD
25% LT TREASURIES VUSTX, TLT, or EDV
25% ST TREASURIES VFISX how about VFSTX

[craigr]

I'm unclear, but I think you're posting returns of the PP with only changing the stock allocation from TSM to slice and dice, correct?

[grayfox]

Yes, it was that Harry vs. Larry study from Trev H on the big HB thread.

Here you go..

I only included the Lumper (TSM,Tot Intl) & Larry (SV,ISV) equity comparisons.

The SUB&H combo of LB,SV,ILV,ISB would have performed in-between, a bit closer to Larry than Lumper.




Considering the other diversification factors in the mix, going all SV,ISV vs Lumping made very little difference in volatility, but nice difference in return.

==

[craigr]

Ah yes. The battle rages on. My usual arguments:

1) Those value returns were done with reconstructed data from indices that didn't exist during the time period and were constructed with hindsight.

2) There are higher costs involved in these value funds in the form of expense ratios, tax costs and in some cases advisor fees on assets under management.

3) Having multiple funds means more rebalancing is necessary and this adds on capital gain tax costs during the transactions.

4) There were long stretches of time where TSM beat value investors (like 1979-1999 - at least 20 years) so you'd have to be a patient person to deal with the market tracking error. This could happen again going forward. There is no guarantee that value will outperform on your time table.

5) This stuff had been tried before and was abandoned.*

6) The primary diversification of the permanent portfolio allocation comes from the stock/bond/cash/gold split and not messing around with the stock allocation.

7) History doesn't repeat. TSM maximizes your after-cost returns and is guaranteed money in your pocket in savings. Value tilting is a gamble that your higher costs are going to pay off in higher returns which may not happen.

BUT, if you just have to value tilt and want to take on the risk then you should do it with the cheapest index funds you can find. Preferably from a company that's been around a while and has a track record of following their stated index. While I don't do this myself as I see little to gain from the strategy, I don't think you're going to do a lot of damage at least compared to other asset allocations that take on far more risk.



* John Chandler, Harry Browne's longtime publisher and business partner, said they dropped the idea of holding "aggressive" funds in the portfolio because they found that whenever these strategies were uncovered they would quickly lose their effectiveness as the market arbitraged the opportunity away. With the wide availability of value indexes today, it is not unreasonable to assume that this could happen with value tilting going forward.

[grayfox]

4) There were long stretches of time where TSM beat value investors (like 1979-1999 - at least 20 years) so you'd have to be a patient person to deal with the market tracking error. This could happen again going forward. There is no guarantee that value will outperform on your time table.

From the 40-year table, it looks the traditional HBPP with TSM beat L-B with Value tilt in only 10 years, or 25% of the years. And five of those years was 1995-1999 the years of irrational exuberance and stock market bubble.

The other years when the traditional HB won were 1973, 1986, 1990-91 and 2007.

So the value tilt seems to be persistent. And there is a lot of academic work that says there is a small and value premium. Those Fama and French guys.

[larryswedroe]

craig
Few points, First you are wrong on TSM beating value from 79-99 as the value premium was 1.8 percent, just less than historical

Second, the difference in expenses of the DFA value funds vs TSM is not all that great especially after subtracting the differences in securities lending and won't matter that much when you have such a low equity allocation. Example-take a 25% equity allocation and add say 20bp in expenses and you get 5bp extra costs for the portfolio.

Third, rebalancing may or may not add much costs as depends on location and use of dividends and new cash to rebalance and the biggest rebalancing is between stocks and bonds. This does add some costs but likely relatively minor

Fourth agree that there is no guarantee of course or there would be no risk. But which is more likely to deliver the superior result is the question and which takes less tail risk--the downside pain is far greater than the upside benefit for most people. So that too should be considered as part of the risks of any strategy.

[grayfox]

BTW, I a set up a poll to name this wonderful new portfolio

Play Name That Portfolio!
viewtopic.php?t=39086&highlight=

[Trev H]

Craig said...

==
There were long stretches of time where TSM beat value investors (like 1979-1999 - at least 20 years)
==

And Taylor often states the same or similar thing spouting 15 years of underperformance by SV vs TSM.


1979-1999 what actually happened...


Showing EHCBL (Extreme Home Country Bias'd Lumper - aka TSM), GCPL (Global Couch Potato Lumper) vs US SV and Global SV.





Where is that 15 or 20 years that Taylor and Craig spout about ?

I only see 1 year (the PEAK of the BUBBLE) where TSM managed to raise it's Bubbly Head above the others. Don't stop there though - follow TSM and the others thru to today to get the rest of the story.

Craig and Taylor have to stop in 1999 (notice it is their only Happy time) in the span, performance wise, compared to the others.

What really gets me about these Lumper comments is how they introduce all kinds of "fears" when trying to convince EVERYONE ELSE "not to diversify".

Yet they "fearlesly" trudge forward with 80-100% of their equity allocation LUMPED into US Large Caps (feeling safe with that).

===

[Trev H]

Greyfox asked..

===
That sounds like a good approach. What would the Vanguard funds be?

6-1/4% LB VTSMX
6-1/4% SV VISVX
6-1/4% ILV VGTSX or VFWIX or VTRIX
6-1/4% ISC VINEX or VSS
===

If using Vanguard Funds and have room in Tax Friendly locations for Small/Value tilts could get the SUB&H (Simplified Ultimate Buy & Hold) equity mix with:

25% Large Cap Index, 500 Index (could use TSM but I would not).
25% Small Value Index
25% International Value (ILV which includes 19.1% EM LV).
25% FTSE X-US Intl Small Cap (ISB which includes 19.9% EM Small)

If you wanted to exclude the low cost managed fund International Value, then you could go with FTSE X-US Intl Large Cap or Total Intl (although would give up the Value tilt).

If Tax Efficiency is a must (mostly taxable, high tax bracket) you might consider this mix of ETF's instead.

25% VV for US Large Market
25% IJS for US Small Value
20% EFV for ILV Developed Only
10% VWO for EM Large Market
20% VSS for ISB (Vanguards FTSE X-US Intl Small Cap ETF).

===

[james22]

Could you graph this, Trev?

25% SV (US/I/EM)
25% CCF
25% TIPS
25% IT

Thanks.

[newbie001]

"Where is that 15 or 20 years that Taylor and Craig spout about ?

Craig and Taylor have to stop in 1999 (notice it is their only Happy time) in the span, performance wise, compared to the others.

What really gets me about these Lumper comments is how they introduce all kinds of "fears" when trying to convince EVERYONE ELSE "not to diversify".

Yet they "fearlesly" trudge forward with 80-100% of their equity allocation LUMPED into US Large Caps (feeling safe with that)."


I don't have a dog in this fight about how to split the stock portion of the PP, but I can't help noticing from this thread and the PP thread that you seem to have a lot of pent-up anger toward so-called Lumpers. It's not life or death, it's marginal improvements that are at stake.

[Trev H]

Newbie...

It is the mis-representation of the data that gripes me.

What Craig and Taylor say (how they say it) is simply not true.

And using that peak of the bubble return data for TSM and specific starting points is simply data mining. Not really useful to anyone, and potentially harmful to many.

You can call it anger if you want, but it is more like frustrating, especially considering they know it is not the actual truth (have been told that in the past by me and others) just like Larry did in this thread.

It will be used agian - and I will continue to call them Lumpers - after all Bernstein coined the phrase (or at least 4 Pillars was where I first heard it).

===

[Trev H]

James22,

Will try to get to that chart soon. Perhaps sometime after work today.

I remember looking at mixes including CCF & REIT in the past (mixed with Stocks/Bonds that did very similar to PP.

==

[Trev H]

James22 asked..

==
Could you graph this, Trev?

25% SV (US/I/EM)
25% CCF
25% TIPS
25% IT
==

I did 3 different mixes below in each using (ITT/TIPS) for the bond allocation and (SV/ISV) for the Equity. In one used 25% Commodities, another 25% Gold, and another 12.5% each Commodities/Gold.

===

[Trev H]

Below is another mix/comparison.

Looking at 50% Larry Equity (SV/ISV) + 50% ShortTerm Treasury

Then reducing the Larry Equity while adding in 10% slices of Gold, keeping the ST Treasury allocation at 50%.

[james22]

Thanks, Trev. That's interesting.

[WileECoyote]

First off let me say that I believe there has been a small cap equity premium and value has slightly outperformed over time as well. Could you run the numbers to see how much of the small cap value premium was from outperformance of small cap vs outperformance of value? Ie was it more size or style that drove returns or truly a combination of the two? I do however realize that the data could be skewed as it is a pieced together time series, while not perfect it could still be instructive.

I also believe the argument that market premiums are often eroded as more players enter the space, it's just the unfortunate fact of competition. So the premium and outsized returns that existed before may not be as wide as they were in the past?

[craigr]

Craig and Taylor have to stop in 1999 (notice it is their only Happy time) in the span, performance wise, compared to the others.

First of all, I think Taylor is a fine gentleman and I'm proud to be "lumped" together with his wisdom.

Secondly, you are inferring a tremendous amount about the future from past historical data that was:

1) From an index that didn't exist most of the time period in question.
2) From data that that was put together to maximize performance in hindsight.
3) Probably simulated from reconstructed historical records.

What really gets me about these Lumper comments is how they introduce all kinds of "fears" when trying to convince EVERYONE ELSE "not to diversify".

You are mis-representing my statements. I (and Taylor) never advocated people not diversify. In my case I just think your primary diversification comes from your stock/bond/cash/gold split and not worrying about dicing the stocks into many pieces. Taylor typically advocates a split between stocks and a high-quality bond fund like Total Bond Market and/or TIPS. Neither of us advocates not being diversified.

The main thrust of your diversification seems to me maximum profits based on past performance of the stock market. I focus on diversification that offers protection in a variety of economic climates but may give up some upside potential if we have a big bull market in stocks. We just disagree on what benefits we should achieve from diversification (you want performance, I want safety).

Yet they "fearlesly" trudge forward with 80-100% of their equity allocation LUMPED into US Large Caps (feeling safe with that).

I sleep like a baby with my portfolio. Yet last year, a multi-factor asset class investor who was sold the bill of goods that stocks alone can offer "diversification" got creamed. Therefore, I'll be my own counsel on what diversification works for what I require.

[Taylor Larimore]

Bogleheads:

What Craig and Taylor say (how they say it) is simply not true.

I'll ignore the innuendo.

"Truth" is not confined to one lay person:

"Most people might just as well buy a share of the whole market, which pools all the information, than delude themselves into thinking they know something the market doesn't." Nobel Laurete, Merton Miller

"It is the easiest thing in the world to look backward on paper and pick out strategies that have won. If you can't do that you are, for the want of a better word, a moron." -- Jack Bogle

From 1937 - May 2006 (date of publishing Random Walk), the average annual return of growth stocks was 10.62%; value, 10.59%. A virtual tie over 69 years. -- Burton Malkiel, page 265

"Whatever small value premium existed seven years ago, there is none now. In fact, going forward, small value premium may even be negative. TSM is a better way to go, even before you start considering taxes. -- Bill Bernstein

"Whether you decide to tilt towards value depends on whether you are willing to bear the associate risk. The market portfolio is always efficient. For most people, the market portfolio is the most sensible decision." Dr. Eugene Fama

"After considering the evidence, Vanguard maintains that investors should generally hold market-wieghted equity portfolios. Even if the small-cap value premium was actionable, investors would face increased risks and cost by tilting their portfolios away from broad-market weightings."--Vanguard

"Much of the early data on small caps is terribly flawed." Jack Bogle

Small Value is risky. During 1998-1999 when the S&P 5900 gained +55.6%, small value lost -7.9%

"We find no evidence that value firms earn higher returns than growth firms. Among mutual funds holding small-cap firms, we document average annual returns of 14.10% for value funds compared to 14.52% for growth funds." --Houge Loughran study (below)

Do Investors Capture the Value Premium?

"You take a big risk in overweighting styles or market caps. You're esssentially betting the market weight is wrong, and I don't know anyone who's been successful betting against the market over the long term." Jack Bogle

"The closer you come to holding the entire market portfolio, the higher the expected return for the risk you take." Nobel Laurete, Wm Sharpe

The Telltale Charts

Investing in Total Market

"Using past performance is such a lousy idea that mutual fund companies are required by law to tell you it is a lousy idea." (Bill Schultheis)

[juhrio]

people usually dump the small stocks in a big downturn like we just had so I would bet that small caps are back to their old premium. I would be for a lot of strategies the old premiums are back.

look at the carnage in the gold junior sector with regards to small stocks. same for small cap commodities stock.

[juhrio]

First off let me say that I believe there has been a small cap equity premium and value has slightly outperformed over time as well. Could you run the numbers to see how much of the small cap value premium was from outperformance of small cap vs outperformance of value? Ie was it more size or style that drove returns or truly a combination of the two? I do however realize that the data could be skewed as it is a pieced together time series, while not perfect it could still be instructive.

I also believe the argument that market premiums are often eroded as more players enter the space, it's just the unfortunate fact of competition. So the premium and outsized returns that existed before may not be as wide as they were in the past?

usually when the premium is finally out is when the strategy will tank. people get out and then the strategy works again after a couple of years. notice when the TSM caught(1999) up the market then tanked a year or so later.

[matt]

Trev H,
I have just one question for you, in two parts:

What was your personal performance in 2008 and did the loss
exceed any historical precedent derived from your backtesting (excluding the Great Depression, which, of course, we all assume can't happen again)?

I think this will answer the question of whether a backtest may be worth betting on going forward.

[Roy]

Bogle is a great man who changed investing for the common person. But he made mistakes too. Thankfully, these have been mentioned for years on several forums by Larry Swedroe:

From Bogleheads:

"As to Bogle on value. Bogle made the incredible error of not looking at asset classes of size and value but of mutual funds that invest in those asset classes. Then made his statement. That is simply mixing apples and oranges. First active value funds buy growth stocks and vice versa. So you don't have pure funds. So the analysis only showed that actively managed value funds did not outperform active growth funds. And another reason is the value funds simply cannot get themselves to buy the really depressed stocks which have the highest returns. Clearly Bogle was wrong in his statement. The data shows otherwise. I have pointed this out several times so I don't know why this keeps getting cited. "

And at Morningstar:

"First as I stated Bogle was simply wrong. He made mistake. He did not compare growth stocks to value stocks. He looked at growth funds and value funds. They both tend to style drift and so you get similar returns, and also the value funds tend not to really buy true value stocks (they dont want to look stupid buying the real dogs."


The future may indeed not resemble the past (for all we know, the equity premium may vanish entirely), but this applies to total market funds too—their returns may be comparatively worse than other classes than they were before. However, both Swedroe and Bogle have also pointed out that the best returns usually fall to those who can stick to whatever strategy they choose.


Finally, even if using TSM in the Harry Browne Permanent Portfolio, the PP concept is so different in its approach to overall diversification, that I would not liken it in any way to a conventional total market approach.


Roy

[WileECoyote]

It seems to me that there is a natural waxing and waning of large vs small in US equities over time - with periods of outperformance followed by periods of underperformance. So what I had always thought about, but not done yet, was the following idea. Split your US equity in half with very large (say BRLIX) and very small cap (BRSIX), then when you add money or rebalance to your US equity you add to whichever is cheaper or has a higher yield. So you would constantly be buying the cheaper cap size when you add money\rebalance and should then capture a greater return on each investment as you are buying the cheap part of the market. Then as the cap shift happens you will begin reaping rewards on the out of favor size and begin adding to the now out of favor size. Sort of an internal rebalancing within US equities. Any thoughts?

[juhrio]

my only problem is that during the bull market in stocks the strategy really underperformed.

PRPFX lagged from around 1996 to 2008.

[grayfox]

Here is how I am thinking about the great debate between LUMPERS and SLICERS.

Consider diversification as having multiple levels.

Level 1. At the top level you have equities, real assets, LT bonds, ST bonds/cash.

If you want to do the official HB PP then choose TSM for stocks, gold for real assets, 30-year Treasuries for LT and mmmf for cash. The top level of diversification probably gives you like 90% of the diversification effect. You can stop there and be done.

Level 2. Or if you want you can go down one level more in diversification:

Split stocks between U.S. and International.
Real assets could be split among gold, silver, oil, CCFs, real estate, etc.
You could even divide LT between Nominal Treasuries and TIPS if you were so inclined.
Cash could be split among CDs, mmmf, and short-term treasuries or even short term corporates.

These variations are not the official HB PP. maybe they would provide an incremental increase in diversification. Maybe they would make it worse. i don't know.

Level 3. You could add one more level of diversification and split U.S. stocks between growth and value and tilt to value. Or maybe go 100% value. Do the same for foreign stocks.

You could also divide between small and large and tilt to small or go 100% small if you are feeling lucky. Maybe this slicing and tilting provides a little more diversification and/or increases the expected return a little, at least according to backtesting. Maybe it increases the risk. Who knows?

Diminishing Returns. I think as you go further down you reach a point of diminishing returns. Is it worth the extra complexity and added expenses of more funds if you go beyond the first level of diversification? Stopping at the first level, I don't see you missing out on much.

There is something to be said for simplicity. I have seen cases where adding complexity can come back and bite you in the butt.

[grayfox]

I myself go a little into the second level and split between U.S. and International. I also went for a short-term bond fund instead of mmmf. Overall my portfolio is split between nominal bonds and TIPS. Maybe I'll go to Level 3 and divide into value, but I don't really like having too many funds. Plus you can loose Admiral status if you divide things up too much. And a lot of those funds have higher E.R.s

The portion of my portfolio that I devote to the HBPP idea originally looked like this:

VTSMX 25% GLD 25% VUSTX 25% VMMXX 25%

Now it looks like this:

VTSMX 12.5% VGTSX 12.5% GLD 25% VUSTX 25% VFSTX 25%

It is not the official HBPP because I decided to divide stocks between U.S. and International. Also I substituted Short-Term Investment Grade bond fund for cash and use VUSTX instead of TLT or EDV. If I wanted to add value and small maybe I would have:

VTSMX 6.25% VISVX 6.25% VGTSX 6.25% VSS 6.25%
GLD 25%
VUSTX 25%
VFSTX 25%

Now that would be seven funds. Is it worth the trouble? If I went for the Harry-Larry Portfolio it would be back to five funds:

VISVX 12.5% VSS 12.5%
GLD 25%
VUSTX 25%
VFSTX 25%

[Roy]

my only problem is that during the bull market in stocks the strategy really underperformed.

PRPFX lagged from around 1996 to 2008.

Hi, Juhrio,

Performance must be judged by portfolio objective. The objective of the HB PP is to avoid losing big in downturns (just like the "Larry" portfolio). (Note that the HB version has a collective expense ratio far lower than PRPFX.) The PP sacrifices relative underperformance in true "bull" markets to accomplish this. The period you mention included 2 serious downturns. Here how the HB PP did (data taken from Craig's board):


1996 4.9
1997 7.5
1998 10.8
1999 4.2
2000 3.2
2001 0.9
2002 7.0
2003 14.0
2004 6.6
2005 8.1
2006 11.0
2007 12.9
2008 1.9

Note how well the PP did during 2001-2002, and 2008. Considering its objective (and mine), Its worse loss in 37 years—about - 4% with a CAGR of 9.38%. I'll take that sort of "underperformance" any day. Even using Trev's data, above, the PP's objective is well met.

Roy

[Trev H]

And a few more classic lumper replies from CraigR and Taylor.

First of all CraigR you said this...

==
There were long stretches of time where TSM beat value investors (like 1979-1999 - at least 20 years).
==

Please provide the data that you used to form that statement.

Also consider this - - - Was that data from:

1) From an index that didn't exist most of the time period in question ?
2) From data that that was put together to maximize performance (of TSM) in hindsight ?
3) Probably simulated from reconstructed historical records ?


Data is Data - we all use the best data possible.

Why on earth would you think that anyone would manipulate the data to favor SV or TSM ?

The data that I use (which you claim to be so un-reliable, since it does not work out in your favor) is:

TSM = US Cap Weighted Market:
=============================
CRSP Market Decile 1-10 1979-1992
Vanguards Total Stock Market Index Fund 1993-5/31/2009

SV = US Small Value:
====================
Russell 2000 Value Index 1979-1998
Vanguards Small Cap Value Index Fund 1999-5/31/2009

Where the raw index data is used, the ER of Vanguards Index fund is backed out of the return.

What possible benefit could there be for Russell to manipulate the data in favor of SV ? Why would they choose SV over SG ?

===

Please do list the data source for you statement !

You specifically said that TSM beat Value for 20 Years.
Taylor has made the same statement in the past but only for 15 years.

Both are an absolute un-truth and you are confusing investors with statements like that.

I can only conclude that you both do that out of some kind of deep fear of investing in anything other than US Large Caps.

If that is not the Case - Please explain why you make such statements.

20 Years, or 15 Years.

Using Russell/Vanguard data below is what actually happened.

This time I eliminated the Global Diversification and only included the US Market TSM and SV (data sources listed above).



Where is that 15 or 20 years where TSM beat SV ?

If you have some data source that shows otherwise, I would like to see it, study it, make my on conclusions as to it's worthyness.

One thing you can believe me on is this.

I am absolutely not Biased, but I am persuaded by Data.

If the data available supported your views (lumping all your equity in TSM is the best bet) I would be shouting that as LOUD as you or Taylor.

It simply does not.

Not the data that I have seen, which I consider to be from credible sources - Russell, Vanguard.

Either one of you put up the actual data that shows how TSM beat SV for 15 or 20 Years.

And you both say it that way (which is my main gripe).

It simply did not happen like that, and that is a complete misrepresentation of the facts and I don't really care who you are, Taylor, Bogle, CraigR, if you are not telling the truth - you will not get any respect from me.

Nor shoud you from anyone.

The Data Please ???

Thanks

[Trev H]

Matt asked...

===
What was your personal performance in 2008 and did the loss exceed any historical precedent derived from your backtesting (excluding the Great Depression, which, of course, we all assume can't happen again)?
===

I can honestly say I don't know what my return for 2008 was. I never checked and I don't really care. I don't see how a single years return is all that important in the long run.

I guess if you were the type to panic and sell low, how you reacted last year would be very important.

Me - I changed nothing in 2008 (did not even rebalance). I normally do rebalance around year end (once a year) but did not at year end 2008 because had heard that FTSE X-US Intl Small Cap would soon be available.

I bought it during the subscription period and rebalanced then.

If you have not figured out by now, I am very much an optimistic person.

Where others might look at 1973-1974 and focus on the losses, I look at the next 10 years and the incredible gains.

I have shown several times how Lumper vs SUB&H compares in Bear Markets.

Lumper suffers far worse.

That is considering you simply do what Larry suggest - as you take on size/price diversification, you reduce equity risk.

Using my mostly Russell/Vanguard data, if you change from Lumper to S&D with 50% exposure to Small & Value stocks, the risk (measured in volatility) and shown in Bear Market loss is completely eliminated by increasing the bond allocation 10%.

Not sure if that answered your question, but it was the best I could do.

Thanks

[Trev H]

Matt asked...

===
What was your personal performance in 2008 and did the loss exceed any historical precedent derived from your backtesting (excluding the Great Depression, which, of course, we all assume can't happen again)?
===

I can honestly say I don't know what my return for 2008 was. I never checked and I don't really care. I don't see how a single years return is all that important in the long run.

I guess if you were the type to panic and sell low, how you reacted last year would be very important.

Me - I changed nothing in 2008 (did not even rebalance). I normally do rebalance around year end (once a year) but did not at year end 2008 because had heard that FTSE X-US Intl Small Cap would soon be available.

I bought it during the subscription period and rebalanced then.

If you have not figured out by now, I am very much an optimistic person.

Where others might look at 1973-1974 and focus on the losses, I look at the next 10 years and the incredible gains.

I have shown several times how Lumper vs SUB&H compares in Bear Markets.

Lumper suffers far worse.

That is considering you simply do what Larry suggest - as you take on size/price diversification, you reduce equity risk.

Using my mostly Russell/Vanguard data, if you change from Lumper to S&D with 50% exposure to Small & Value stocks, the risk (measured in volatility) and shown in Bear Market loss is completely eliminated by increasing the bond allocation 10%.

Not sure if that answered your question, but it was the best I could do.

Thanks

[Trev H]

Matt,

Below is a chart that I did for the TR Fund thread a day or two ago.

It shows a good example of the Simple/Global 3F Diversified mix vs TR Fund Equity allocations.





If you will notice, the TR Fund at 80/20 and 50/50 end up in practically the same place in 2008 (40 years).

And the example to look at (that sort of addresses your question, I think anyway) the combinatin of

12.5% each LB,SV,ILV,ISB, 50% ITT

vs either of the TR Fund mixes.

Anyway that is a good example of how diversifying the Equity side via Size/Price does not = a more risky portfolio.

By any of the standard measures, StDev, Sharpe, Bear Loss, the 3F diversified equity mix, was much less risky than the Lumper - Home Country Bias'd TR Funds (80/20 or 50/50).

===

[Taylor Larimore]

I don't really care who you are, Taylor, Bogle, CraigR, if you are not telling the truth - you will not get any respect from me.

Nor shoud (sic) you from anyone.

The Data Please ???

Trev:

Please review our Rules of Etiquette:

"We expect this forum to be a place where people can feel comfortable asking questions and where debates and discussions are conducted in civil tones. Respect your debating opponents. Debates are about issues, not people. If you disagree with an idea, go ahead and marshal all your forces against it. But do not confuse ideas with the person posting them; at all times we must conduct ourselves in a respectful manner to other posters."

This is the data you requested:

Annualized returns for periods ending 6-30-00:

Style.......1 Yr....3 Yr.....5Yr....10 Yr...15 Yrs.
LCG......27.19...27.04...24.98..17.85....17.15
LCB........8.93...17.29...20.35..15.60....15.23
LCV.......-5.21....8.74...15.16..13.36....13.45

MCG.....57.24...31.09...24.81..18.14....16.82
MCB.....11.87...12.36...16.09..14.12....14.23
MCV.....-2.56.....7.23...13.20..12.77....12.64

SCG......55.14...24.42...20.86..17.12....15.64
SCB......17.77...10.08...15.30..13.03....12.02
SCV.......3.29.....3.55...12.58..11.80....11.34

Source: Morningstar

http://socialize.morningstar.com/NewSocialize/forums/p/187006/2274458.aspx

"Past performance is no guarantee of future performance."

[juhrio]

my only problem is that during the bull market in stocks the strategy really underperformed.

PRPFX lagged from around 1996 to 2008.

Hi, Juhrio,

Performance must be judged by portfolio objective. The objective of the HB PP is to avoid losing big in downturns (just like the "Larry" portfolio). (Note that the HB version has a collective expense ratio far lower than PRPFX.) The PP sacrifices relative underperformance in true "bull" markets to accomplish this. The period you mention included 2 serious downturns. Here how the HB PP did (data taken from Craig's board):


1996 4.9
1997 7.5
1998 10.8
1999 4.2
2000 3.2
2001 0.9
2002 7.0
2003 14.0
2004 6.6
2005 8.1
2006 11.0
2007 12.9
2008 1.9

Note how well the PP did during 2001-2002, and 2008. Considering its objective (and mine), Its worse loss in 37 years—about - 4% with a CAGR of 9.38%. I'll take that sort of "underperformance" any day. Even using Trev's data, above, the PP's objective is well met.

Roy

my problem is even people who are staunch HBPP may suffer during bull markets. it took almost 9 years of moving sideways through two 50% bear markets to catch up.

downside risk might be easily managed and allow to tilt the portfolio to stocks during a long-term bull market.

I know that is getting away from the program, but it may save a lot of headaches due to underperformance.

[Trev H]

Taylor,

Do you not understand that the 15 year annual return can be changed by that margin by only a few of the most recent years.

And how misleading it can be to twist that around and say TSM outperformed SV for 15 or 20 years.

That simply did not happen.

Large Growth and TSM Bubbled in a 4-5 year span and YES it did skew the 15 year annual return data accordingly.

Not at all the same as "for 15 years".

That part is not true - you are misleading folks.

CraigR is too.

If Small Growth had a 400% return this year and Small Value -100% would it be accurate to say that SG outperformed SV for 15-20 years.

Absolutely not - IT WOULD NOT BE TRUE !!!

What you and CraigR are saying is not true.

You need to word that correctly - that is my Gripe with you and CraigR or any one else that states it that way.

It simply did not happen like that, and you are misleading folks when you say it did.


===

[DP]

Hi,
Some interesting discussions. Re. Lumpers vs. Splitters, I have yet to see any compelling evidence or reason why someone should invest in a Total Market fund which is very much overweighted to Large Cap stocks. While the market is a useful benchmark, I care much more about potential returns (and risk) going forward then I do about tracking error. Not only is there no inherent reason to expect Large Cap stocks to outperform going forward, there are arguably reasons to expect small cap and value to outperform, though it is certain they will not always do so. While it may be argued that SCV is riskier, the portfolio's being discussed limit that risk.

FYI, I very much agree with the concepts of combining PP and Larry's portfolio's. I am tracking my own version in the R48 challenge, also pulling in idea's from Scott Burns:

Equal parts:
SCV, Emerging Markets (or IS), LT Bonds (or intermediate treasuries), Gold, Global Bonds (or ST treasuries), Tips.

Backtest spreadsheet returns, using the primary asset class except ST (2yr) treasuries rather the Global is:
CAGR: 11.6%, StdDev 8.9%, Sharpe .68, Max annual loss: 9% (2008).

Don

[matt]

Trev H wrote:

I can honestly say I don't know what my return for 2008 was. I never checked and I don't really care. I don't see how a single years return is all that important in the long run.

Based on your fondness for backtesting, my assumption is that it was a significant factor in determining your asset allocation. I also assume that most of your thinking and backtesting on what allocation to use came between 2000-2007. I assume that this led you to slice and dice and get as far away from exposure to U.S. Large Growth as possible.

Assuming that the above is reasonably accurate, your investment experience since mid-2007 has probably been quite poor and I would bet that the losses were an outlier that fell far outside the expectations created from backtesting. In other words, this bear market did not fit the historical pattern almost immediately after you identified and sought to take advantage of the pattern.

This brings into doubt whether you can backtest your way into the best asset allocation. I believe that the out-of-sample testing for most backtesters failed over the past 24 months. From slice-and-dicers to Wall Street quants, the past was not prologue. However, Harry Browne's strategy was well within the bounds of expectations; it did not break down in a crisis.

I actually get the feeling that in 1998-99, Trev H would have been the one trying to convince the rest of us to load up on Large Growth because of the impressive backtest results.

Of course, it is all too easy to criticize after the fact. But I have, and still do, believe that an investment strategy must be forward-looking. A backtest is clearly not. Harry Browne's strategy is, even if it acknowledges that the future is unknowable.

My "pre-crisis" thoughts from May 2006 on this topic appear to have been validated; critical thinking works better than backtesting.

Here's my prediction of what is going to happen at some point in the future. We will have a substantial bear market that goes across multiple asset classes. All of the "diversifiers" that everyone is adding to their portfolios right now (value, emerging markets, REIT, commodities, precious metals, etc.) because of wonderful backtesting engines showing theoretical protection in previous bear markets, will provide little help.

We will all look back at this and the folly will be obvious: if everyone is diversifying in the same way, it cannot work because there are fewer investors to take advantage of doing the wrong things. Now that everyone agrees that diversification is the right thing, I believe it will have a serious breakdown during a future financial crisis. At that point, maybe investors will shift back to just the "core" asset classes and diversification will start to work again.

http://socialize.morningstar.com/NewSocialize/forums/p/174205/2164024.aspx

Nothing has changed in the past three years, except, perhaps, for what the backtest is now telling you to do.

As for value investing, I actually favor it quite a bit. However, I vastly prefer Ben Graham's methods over those of Fama and French. I want to utilize value investors that think, not just compute. Anyone can compute, which is exactly why quantitative approaches are at risk of breaking down.

[juhrio]

just ran across this lazy portfolio article.

http://www.marketwatch.com/story/lazy-p ... genumber=2

[Roy]

While it may be argued that SCV is riskier, the portfolio's being discussed limit that risk.

Hi, DP.

This is precisely the point. The problem is it gets continually lost, ignored, or forgotten in the comparative discussions, which view asset classes in isolation (e.g. Growth vs. Value or Large vs. Small). As Larry so often points out, tilting to size and value permits you to lessen the equity allocation—and consequently shorten fat tail risk. I think Trev's data shows the results of this quite nicely.

The HB PP concept almost forces one to see the portfolio as a whole because it has 3 volatile asset classes, which have low correlation to each other, plus a fourth very stable class. Thus, it seems to avoid the "religious wars" of the asset classes and its very nature forces a whole portfolio analysis.

FYI, I very much agree with the concepts of combining PP and Larry's portfolio's. I am tracking my own version in the R48 challenge, also pulling in idea's from Scott Burns:
Don

I think this has merit and Trev has shown some interesting variants that retained the Long-term bonds, Gold, and Short Treasuries. I do get nervous monkeying with the HB PP beyond the equity portion. But the nervousness comes mostly from my seeing it as an extremely successful, if not the most successful, all-weather model.

Roy

[craigr]

Ok I'm not near a computer now so I'm responding on my cell phone. Sorry for any typos.

When I'm talking about diversification I'm speaking as if building a home. I want it built well such that I can live in it safely and for a long time. I under stand that my home may have to survive a cat 5 hurricane one day so I will put money into reinforcing it. I also know the area has an occassional tornado so I put in a storm shelter. For the days that are pleasant I'm going to have my home facing the direction to make best use of the sun. For days that it's rainy I'm going to make sure it has good drainage and isn't on a flood plain. Because the cost of maintaining that home needs to be paid each year, I'll build it with energy efficiency so I don't needlessly waste money. Etc.

I feel it's much more important to worry about the basics first and make sure you can weather the problems that show up and also be able to enjoy the good times.

What Trev is worrying about is what color to paint the rooms. It's so far down on the list of considerations of investors in terms of useful diversification that it just isn't worth worrying over. Get the other stuff straight first. Just because Trevs data says the average hurricane is only a cat 3 does not mean it is optimal to forgo these other protections and try to get higher returns.

Lastly, when you are splitting these assets up into so many pieces you are transitioning from the realm of investing to speculating. You are trying to beat the market. This carries risk no matter what the data says and there is no free lunch in investing. When someone is offering you a free lunch, you are probably going to get food poisoning.

[DP]

Craigr,
When I started reading your analogy I thought at first you were arguing the other side.

For the days that are pleasant I'm going to have my home facing the direction to make best use of the sun.

When I read this I immediately thought that for the times that are prosperous, I want to invest in stocks that have the most potential for growth: undervalued small cap. It would be interesting to see what the data shows is the best style only during bull markets. I very much doubt it would be LCB or LCG.

Again in terms of your analogy, I don't think the question is whether future storms are cat 3 or cat 5 - that's what the limited allocation to equity and diversification into the other specific asset classes is to address. I think the question is what is the best way to position your home to make best use of the sun (what asset class/style is best during prosperous times)?

Don

[craigr]

Arguing through analogy is always a problem.

Fundamentally I think these slice and dice equity debates are always about trying to beat the market. They certainly aren't about diversification because it has been clearly shown (again) that you can't get meaningful diversification by owning stocks alone.

So for me the issue comes down to whether this is a good speculative opportunity. It may be but I have no way to know. But I don't speculate with my core capital so that's why I just stick to broad based index funds and accept the returns the markets make available to everyone.

[Trev H]

Ok - using the Data that Taylor included, it you wanted to look at a specific starting and ending point that favored TSM, that would be it.

Even though it is a really bad example of data mining, misleading, etc - lets take a look at that period, comparing TSM, SV and the SUB&H combo of 50/50 LB/SV.

Taking Taylors data and reducing it to below..

Code: Select all

===
Annualized returns for periods ending 6-30-00:

Style....15 Yrs
LCB.......15.23
SCV.......11.34
===


It did not include the returns of TSM, but LCB is close enough (very little difference in the returns of TSM and LCB). Also note that I do not have the returns (mid year 2000 so can't get to that point, but can show year end 1999, 2000 and thereafter).



Not sure exactly what Data M* was including - Perhaps they were including returns of (all LCB, all SCV funds, including high cost active managed ???) not sure about that.

The data I am including =

TSM = US Cap Weighted Market:
===============================
CRSP Market Decile 1-10 1985-1992
Vanguards Total Stock Market Index Fund 1993-2008

LB = US Large Blend:
========================
Vanguards 500 Index Fund 1985-2008

SV = US Small Value:
=====================
Russell 2000 Value Index 1985-1998
Vanguards Small Cap Value Index Fund 1999-2008

Using Index / Vanguard Fund Data the spread in TSM & SV returns was of course high at the Peak of the LG Bubble. Difference in 15 year CAGR = 4.65

One thing that I want to point out, is that I do not actually suggest that anone go all SV (now Larry might, or at least he does that himself, may not suggest it for anyone else), but my suggestion is that you consider breaking the US Equity into 50/50 LB/SV.

Notice that in 1997 the LB/SV combo and TSM were basically tied.

TSM bubbled and pop'd

50/50 LB/SV did not.

50/50 LB/SV actually suffered very little tracking error compared to TSM, but saved you from the bubbly experience that TSM went thru.

Taylor - I have analized your data, and above are my comments.

My commitment to showing the data results (for quite some time now) has been to show it in the most positive way, most beneficial way, not data mining specific periods to favor one strategy over the other, but showing the full span, the entire truth, including the longest period possible using the best data that can be provided.

Even though I can't say that the data you continue to provied (from that post by Mel 06/30/2000) is absolutely inaccurate. I will say that it is not helpful, if anything it is harmful, and should not be used (not the way you use it). It is simply a extreme example of data mining to try and convince others to invest in TSM only.

I would make on exception to that - it would be completely appropriate to show that as long as you also show the 15 year annual returns for TSM vs SV from 1974-1988 (a period that strongly favored SV 20.94% over TSM 12.34%).

Would like to hear your comments on the suggestions above. Specifically if you think your data is helpful or harmful (the way you present it) and if you will continue to post that going forward without a counter balance showing TSM & SV 1974-1988.

I know that your intentions are to be helpful here...

If you do ever post that again, without the counterbalance, I will begin to question that.

I would also like to address CraigR's statements, which I do feel are absolutely incorrect (TSM outperforming SV for 20 years) from 1979-1999. If he will provide the data source as you did - I will be glad to do that.

===