Asset allocation tutorial?

[LOL!]

Here is a very recent paper by Carl Isrealsen "The Benefits of Low Correlation" which studies everal asset allocation models of a portfolio in withdrawal mode (i.e. a retiree's portfolio). If this study does not turn you into an asset allocator, I don't think anything will

Summary

There are several quantifiable benefits of lowering the correlation of a retirement-withdrawal-mode portfolio's component assets. First, there is a dramatic reduction in the volatility of the portfolio's performance (i.e., lower standard deviation of return). Second, there is a significant reduction in the worst-case portfolio loss, or maximum drawdown. Third, the likelihood (or frequency) of loss is minimized. Fourth, performance does not suffer if sufficient diversification is achieved.

This study suggests that maximum portfolio loss, frequency of loss and probability of recovery following a loss are quantifiable measures of the benefits of low correlation. Furthermore, this study suggests that these three measures may have greater intuitive appeal to investors than standard deviation of return.

As was shown by the mathematics of recovery for portfolios in withdrawal mode, avoiding large losses is of paramount importance. Achieving low correlation among the assets in a portfolio—the key to generating a multitude of benefits—requires the use of a variety of low-correlated assets. Some of the needed assets may not fit the standard paradigm of a traditional retirement portfolio, namely commodities and REITs.

 

[sunshine]

I have no taxable accounts...yet

 

[LOL!]

Thanks everyone for your responses and page "views".

What about fixed income investments, what to consider when choosing (i.e. diversifying) among TIPS, treasuries, munis corporate as well as long term, intermediate term, short term durations.

We have so far written about and discussed allocation among the equity classes. I would like to put it at least a little bit about fixed income asset allocation. I am less clear about this than I am about equity allocation, so I apologize in advance if I muddy the waters for you. I am not going to go into discussion about how inflation & interest rates affect bonds or the definitions of maturity, duration, yield-to-maturity, TIPS, etc.

Some experts suggest that one should take their risk on the equity side of things and use the safest investments for the fixed income side. Also, just as with equities one can do total market weights on the fixed income side or one can slice-and-dice.

For a total market weight, one can go with something like the Vanguard Total Bond Market index fund VBMFX or its etf BND. Some folks do not like this because it contains mortgage-backed securities (MBS). The same folks also do not like a total market weighted bond index because they believe that one should avoid long bond maturities and durations because the one is not rewarded for taking the risks associated with these bonds.

The slice-and-dice folks suggest that one stick with short-term bonds, intermediate bonds and TIPS. Some discussions about all this on the Diehards forum:
Bogleheads :: View topic - Total Bond Index v. Intermediate-Bond Index
Bogleheads :: View topic - Fixed income portion of allocation - TIPS, GNMA fund, bond?
Bogleheads :: View topic - Questions for Larry on TIPS
Bogleheads :: View topic - Why treasuries? Why not Short or Intermediate Invest Grade?
Bogleheads :: View topic - Larry's rely to "Take profit in TIPS" on M*

What I found intrigueing were the messages by Larry Swedroe who really likes TIPS (Treasury inflation protected securities). He likes to buy/hold them when they are yielding above 2%-2.5% (plus the inflation component) and likes to sell or cut back on them when the yields go below 2%. So you can do a little market timing in your tax-deferred portfolio. (Note that VISPX and TIP have gone up 10% in the last 6 months).

I know I've left this hanging, so please join in the discussion and post links.

 

[RetireeRobert]

Link to wikipedia article on asset allocation, Brinson study, Janhke, and more. And just when I thought I was starting to understand........

Asset allocation - Wikipedia, the free encyclopedia

 

[RetireeRobert]

Here is a very recent paper by Carl Isrealsen "The Benefits of Low Correlation" which studies everal asset allocation models of a portfolio in withdrawal mode (i.e. a retiree's portfolio). If this study does not turn you into an asset allocator, I don't think anything will

Excellent article. Thanks for this link. Interesting data on 2-class, 3-class, all the way up to 7-class asset mixes and attendant results.

One thing that still bugs me is this study, like Merriman, takes data from 1970-2006. Possibly because for some classes, homogenous earlier data may not have been available.

If some of these dudes would do a study at least for US equities, bonds, cash in varying proportions from say 1926-2006, I would be a happy camper.

 

[jIMOh]

Excellent article. Thanks for this link. Interesting data on 2-class, 3-class, all the way up to 7-class asset mixes and attendant results.

One thing that still bugs me is this study, like Merriman, takes data from 1970-2006. Possibly because for some classes, homogenous earlier data may not have been available.

If some of these dudes would do a study at least for US equities, bonds, cash in varying proportions from say 1926-2006, I would be a happy camper.

It is a good article.

Some anecdotal comments:

first portfolio was 100% equites, and they concluded it was too volatile during draw down.

second portfolio was 100% equities and they still concluded it was too volatile during draw down.

third porfolio was 75-25 and they commented on the remarkable improvement over previous examples

fourth portfolio was 60-40 (60-20-20) and they commented that this also worked well during draw down.

They did their assignment of presenting the 7 asset draw down model, but keeping all 7 "equal", IMO, is not realistic. I could see 35-35-6-6-6-6-6 as a possible break down- they even commented at beginning a 40-60 portfolio worked much better for draw down that the 60-40 portfolio.

The end conclusion that the goal is a less volatile portfolio during draw down makes lots of sense. How the authors reached that conclusion can be questioned, unless they provide a tool to vary the weightings of asset classes (to suggest only 14% equities for draw down makes little sense to me).

 

[W2R]

Yes, that was a very interesting article. It was especially interesting to see what a difference adding assets like cash and commodities made. I like cash, but I have never really considered commodities at all.

This is a terrific thread! Hence the bump. There is a lot to digest in it.

 

[RetireeRobert]

See Index Funds | DFA Funds Approved Advisor - Dimensional Fund Advisors Approved but many folks would say that these folks did some fast moves with the numbers.

Looked at what they had for various allocations over 25, 35, 50, and 80 periods.

Interesting that for any allocation with any sizeable percent in equities, the standard deviation for the 80 year period was way larger than for any of the shorter periods.

Guess the Great Depression of the 30's had quite an effect still skewing results 80 years later.

 

[LOL!]

We have covered equities and fixed income. Are there some other asset classes that we should be concerned with? The short answer is YES. There are some asset class have give a good return and have a low correlation with equities and with fixed income. (There is no point in owning an asset class that has a negative long term return even if it is not correlated with stocks and bonds!)

We have already hinted about some of these asset classes. Take a look at this paper in the Journal of Financial Planning which shows the correlation of 18 different asset classes to each other. Real estate is an asset class with a good return and relatively low correlation to both bonds and stocks.

There are several ways to invest in real estate. One is to directly buy real estate. There are many landlords on this forum. There are many ex-landlords on this forum as well. One can search the forum for pros and cons of owning investment real estate. Another way is to own shares in a REIT or real estate investment trust.

Folks in the total market camp state they do not need to own REITs as a separate asset class since by owning the total market, they own REITs anyways. To see the amount of REITs in a total market fund, you might read this recent Diehard post. It appears that a significant fraction of a small cap value index fund is REITs.

The slice-and-dice folks like to have REITs as 5% to 10% or more of their portfolio. The question comes up: do you count REITs as part of your equities or something separate? It's debatable, but you should note that Morningstar does not break out REITs for you. There are many REIT index funds and ETFs to choose from if you go this route.

Another possibility is the TIAA real estate annuity. For folks who have TIAA-CREF as a 403(b), the TIAA real estate is somewhat different from a REIT since it owns real commercial estate with less leverage than a typical REIT. This fund has quite a track record of consistent positive returns for many years.

In any event, REIT dividends are taxable as ordinay income, so one should avoid REITs in taxable accounts and hold them in a tax-advantaged account like a 401(k), IRA, or Roth IRA.

Any comments about REITs? Do you believe in using them?

Next up: Commodities and natural resources.
Later: Rebalancing.

 

[RetireeRobert]

Real estate is an asset class with a good return and relatively low correlation to both bonds and stocks.

There are several ways to invest in real estate. One is to directly buy real estate. There are many landlords on this forum. There are many ex-landlords on this forum as well. One can search the forum for pros and cons of owning investment real estate. Another way is to own shares in a REIT or real estate investment trust...........

.........There are many REIT index funds and ETFs to choose from if you go this route...............

Another possibility is the TIAA real estate annuity. For folks who have TIAA-CREF as a 403(b), the TIAA real estate is somewhat different from a REIT since it owns real commercial estate with less leverage than a typical REIT. This fund has quite a track record of consistent positive returns for many years.

Barclay's iShares has 5 new US real estate ETF's and a new International (ex-US) real estate ETF. Expense ratios on all six of these ETF's is .48%--not bad.

Symbols are: FIO---Industrial/office
REM---Mortgage Reits
FTY---Real Estate 50
REZ---Residential
RTL---Retail

And the International (ex-Us) World Property---WPS
I think they mentioned something about world property subclass ETFs too--Europe, Asia, but I can't find the info on that now.

Go to iShares ETF for US investors - Exchange Traded Funds for more info on their ETF holdings.

 

[bamsphd]

What price diversification?

When I was trying to find a good fund or ETF for a particular asset class, I found the info at Eric Haas's altruistfa.com site helpful. Here is a link on the foreign small cap asset class: Foreign Dev Mkts Sml Cap Funds but there is lots more info on the site.

I love the altruistfa.com link LOL. Sorry I'm behind on my homework. I'll get it done Real Soon Now....

I've been a Vanguard index fund based slice-and-dicer since graduate school. I'm also extremely cheap when it comes to expense ratios. So my current worst expense ratio investment is 0.30% for Vanguard's Admiral Emerging Market fund.

There are two obvious asset allocations where I don't have any significant exposure. One is international small cap stocks, the other is commodities.

Based on http://www.altruistfa.com/Intlsmallcapfunds.htm the cheapest index based international small-cap option is IFSM at 0.50%. (I don't want a "managed" option, even from Vanguard) I'm a bit tempted by IFSM, and even asked them to send me the prospectus. Though I'm having difficulty tolerating the idea of paying 0.50%.

As I already knew, and http://www.altruistfa.com/commodityfunds.htm confirms, the cheapest commodity option is PCRIX at 0.74%. Though it would clearly have been a nice investment over the past few years, I know I will not convince myself to invest in a 0.74% ER fund, so no commodities for my portfolio yet.

Does anyone have data suggesting that I'm being penny wise and pound foolish for avoiding these sectors based on their investment costs? Where do you draw the line on expenses, and how did you figure out where that line should be?

 

[lswswein]

Does anyone have data suggesting that I'm being penny wise and pound foolish for avoiding these sectors based on their investment costs? Where do you draw the line on expenses, and how did you figure out where that line should be?

This is a good question which is not easy to answer. It is a function of your current asset allocation and the characteristics of the new asset class you are adding. Eric Hass the guy behind altruistfa has a paper on it
http://www.altruistfa.com/Haas%20Expense%20Ratio%20Study%20working%20paper.pdf

There are a lot assumptions in the paper. One way I would suggest you can look at it is to simulate a portfolio with and without Commodities and look at the final returns.

-h
p.s: I do have PCRIX at 0.74 for 5% of the portfolio. I also try to keep my total ER down but decided it was worth it. Please do your own due diligence. Also the couple of yrs Commodities have been a rage - maybe wait a couple of yrs you might see more options and lower ER. Trade off between that and any oppurtunity costs

 

[lswswein]

Btw you can you this spreadsheet to do the simulations to compare results with & without commodities. Also remember this is all past performance - after a point it is a leap of faith because of GIGO.

Simba Backtest

-h

 

[jIMOh]

This is a good question which is not easy to answer. It is a function of your current asset allocation and the characteristics of the new asset class you are adding. Eric Hass the guy behind altruistfa has a paper on it
http://www.altruistfa.com/Haas Expense Ratio Study working paper.pdf

There are a lot assumptions in the paper. One way I would suggest you can look at it is to simulate a portfolio with and without Commodities and look at the final returns.

-h
p.s: I do have PCRIX at 0.74 for 5% of the portfolio. I also try to keep my total ER down but decided it was worth it. Please do your own due diligence. Also the couple of yrs Commodities have been a rage - maybe wait a couple of yrs you might see more options and lower ER. Trade off between that and any oppurtunity costs

The goal is to make money, not minimize expenses.

If you and I each invested 10k two differnt funds, and my ER was 1% and returned me 10% and you ER was .1% and returned you 5%, which is better?

Me: 10k turned into $11,000. I paid $100 in expenses.
You 10k turned into $10,500. You paid $10 in expenses.

You paid less and earned less. I paid more and earned more. What is better?

If a fund returns you 5%, the expenses have already been accounted for in the 5% it returned to you.

The question you need to ask is can you find something which invests in the same exact thing for a cheaper cost, in that case the other fund should have a return which is higher because of the lower expenses.

 

[lswswein]

The goal is to make money, not minimize expenses.

The question you need to ask is can you find something which invests in the same exact thing for a cheaper cost, in that case the other fund should have a return which is higher because of the lower expenses.

I do understand that. The reason to look at ER so closely is because of the simple fact. ER are guaranteed, while returns are not. So you minimize what you can. The previous poster has done his home work and PCRIX is the cheapest fund for the asset class he is looking for. The question is does it provide enough diversification to his portfolio to take a ER hit.

-h
p.s: I guess the way to approach it is given a asset class characteristics what is the ER that makes it ok to add a fund in that asset class. Then you see if any fund like that exists. You could also do a study of given an ER is there any fund you could add that adds to diversification. Anyway this is all past performance so should be taken with a spoonful of salt

 

[ziggy29]

The goal is to make money, not minimize expenses.

If you and I each invested 10k two differnt funds, and my ER was 1% and returned me 10% and you ER was .1% and returned you 5%, which is better?

Me: 10k turned into $11,000. I paid $100 in expenses.
You 10k turned into $10,500. You paid $10 in expenses.

You paid less and earned less. I paid more and earned more. What is better.

What was better historically? The higher fees and higher return.

What will be better looking into the future? Who knows? If we have enough empirical and statistical evidence that the higher fees lead to higher returns, then maybe it is better to keep using the more "expensive" fund. But if we think the past may have been random chance or dumb luck? Maybe in that case it would be best to choose the lower expenses.

The real problem is that we don't know the right answer because we can't really determine whether or not we can reasonably and statistically expect outperformance by the high-fee fund going forward.

 

[jIMOh]

I do understand that. The reason to look at ER so closely is because of the simple fact. ER are guaranteed, while returns are not. So you minimize what you can. The previous poster has done his home work and PCRIX is the cheapest fund for the asset class he is looking for. The question is does it provide enough diversification to his portfolio to take a ER hit.

I don't look for guarantees when investing- that is what CDs are for. I look for probabilities and to manage risks. Returns take care of themselves from there. Positive returns probably.

Generally speaking, most people would not knowingly invest in something which would lose money. Many people will probably invest in something if it would probably make them money. Depends on the risk.

I agree with the point that finding a fund in a new asset class which will probably have positive returns is a good thing. The risk of adding the fund is weighed against the likelihood of seeing improved long term returns or reduced volatility short term.

However choose low expense funds is having the expense tail waive the return dog. If two funds invest in the same exact thing, the one with lower expenses will have higher returns.

If two funds invest in similar things, it's possible one approach might be better than the other, so there is more to this than looking at expenses. For example using valuation weighted index vs cap weighted indexing. Both indexes might hold same securities, but one weighting might return better long term than the other.

So if the fund is managed, I would weigh the PHILOSOPHY of the fund higher than the expenses, assuming the fund is expecting to provide positive returns to the portfolio.

That is my opinion and I am sure there are other opinions as well, some of which have already been discussed.

 

[jIMOh]

What was better historically? The higher fees and higher return.

What will be better looking into the future? Who knows? If we have enough empirical and statistical evidence that the higher fees lead to higher returns, then maybe it is better to keep using the more "expensive" fund. But if we think the past may have been random chance or dumb luck? Maybe in that case it would be best to choose the lower expenses.

The real problem is that we don't know the right answer because we can't really determine whether or not we can reasonably and statistically expect outperformance by the high-fee fund going forward.

If we are talking about two different investments (not similar ones, but different ones), we look at total return. Any other measure could be luck or lead to apples to pigeons analogies.

In addition if expenses were the driving factor, why invest in bonds, as MOST bond funds have higher expense ratios than most equity funds (very general statement, but bonds are more expensive to trade, so expenses are higher). Even bond index funds have higher expenses than equity index funds- so expenses are not the driver when choosing an investment in this regard.

A factor, but not the driver.

If the example I presented was two funds investing in similar securities, then I would still look at philosophies of the two funds before expenses, because a 5% difference in return is too much to pass up, to me, for the risks I assume while investing.

One example I like to use is PRFPX. Not the cheapest fund around, but I have yet to find a fund which invests in foreign currencies and US currencies at same time, for example, plus commodities, plus stocks. Good diversifier.

 

[bamsphd]

This is a good question which is not easy to answer. It is a function of your current asset allocation and the characteristics of the new asset class you are adding. Eric Hass the guy behind altruistfa has a paper on it
http://www.altruistfa.com/Haas Expense Ratio Study working paper.pdf

This is EXACTLY what I was hoping to find. Thanks.

 

[bamsphd]

The goal is to make money, not minimize expenses.

Of course. I wouldn't be in my "expensive" 0.30% emerging market fund if I was simply minimizing expenses.

However, I also know that I can not pick the best performing investments in advance. If I could, I would be running a hedge fund and collecting millions of dollars a year in fees. I wouldn't be asking for investment advice here!

If you and I each invested 10k two differnt funds, and my ER was 1% and returned me 10% and you ER was .1% and returned you 5%, which is better?

Me: 10k turned into $11,000. I paid $100 in expenses.
You 10k turned into $10,500. You paid $10 in expenses.

Assuming both of the underlying investments have the exact same risk. Then over the long term, both underlying investments should have the same return before expenses. If you assume that, do you still prefer a 1% ER over a .1% ER for a long-term investment? I believe that academic research supports the prediction that your return will be reduced by your ER. See http://www.altruistfa.com/Haas Expense Ratio Study working paper.pdf footnote 10 for starting points.

Would you still like the investment at a 4% ER? Personally, I assume I can only withdraw about 4% before expenses, so if I am paying a 4% ER then my safe withdrawal in 0%.

 

[ats5g]

We have covered equities and fixed income. Are there some other asset classes that we should be concerned with? The short answer is YES. There are some asset class have give a good return and have a low correlation with equities and with fixed income. (There is no point in owning an asset class that has a negative long term return even if it is not correlated with stocks and bonds!)

We have already hinted about some of these asset classes. Take a look at this paper in the Journal of Financial Planning which shows the correlation of 18 different asset classes to each other. Real estate is an asset class with a good return and relatively low correlation to both bonds and stocks.

There are several ways to invest in real estate. One is to directly buy real estate. There are many landlords on this forum. There are many ex-landlords on this forum as well. One can search the forum for pros and cons of owning investment real estate. Another way is to own shares in a REIT or real estate investment trust.

Folks in the total market camp state they do not need to own REITs as a separate asset class since by owning the total market, they own REITs anyways. To see the amount of REITs in a total market fund, you might read this recent Diehard post. It appears that a significant fraction of a small cap value index fund is REITs.

The slice-and-dice folks like to have REITs as 5% to 10% or more of their portfolio. The question comes up: do you count REITs as part of your equities or something separate? It's debatable, but you should note that Morningstar does not break out REITs for you. There are many REIT index funds and ETFs to choose from if you go this route.

Another possibility is the TIAA real estate annuity. For folks who have TIAA-CREF as a 403(b), the TIAA real estate is somewhat different from a REIT since it owns real commercial estate with less leverage than a typical REIT. This fund has quite a track record of consistent positive returns for many years.

In any event, REIT dividends are taxable as ordinay income, so one should avoid REITs in taxable accounts and hold them in a tax-advantaged account like a 401(k), IRA, or Roth IRA.

Any comments about REITs? Do you believe in using them?

Next up: Commodities and natural resources.
Later: Rebalancing.

Well, both Swenson and Malkiel believe in using REITs for diversification. Barclays wrote an article Real Estate Investing the REIT way which concluded:

Once the data was cleansed for these biases, the study shows that, statistically, the risk and return attributes of REITs are not dissimilar to those of direct real estate. In short, REITs can be viewed as an effective proxy for direct real estate investing.

This simple yet important conclusion has profound implications for investors. Namely, it suggests that REITS may be used as both a substitute as well as a complement for direct real estate investing. For example, the immediacy of market exposure afforded by REITs can be employed as a temporary vehicle while an investor seeks out the next shopping mall or office building. The use of REITs as a tactical asset allocation vehicle is a sensible solution that plugs any potential misweight to a strategicreal estate exposure...

Also, a 2004 paper by Joe Gyourko at Wharton came to a similar conclusion, that despite differences in valuation, liquidity, leverage and fees, REITs may be considered real estate. [Gyourko, J., Real Estate Returns in the Public and Private Markets: A Reexamination Following the Rise of Equity REITs, 2004, The TIAA-CREF Institute.]

btw - the reason that the TIAA account hasn't had a down year is that investment properties [as measured by NCREIF] haven't had a down year since the early 1990's.

Allocation to REITs is probably good, but I wouldn't go hog wild. Not too mention, that all those peeps who are tilting towards small/value [via VISVX] already have a healthy dose of REITs.

- Alec

 

[bots2019]

all those peeps who are tilting towards small/value [via VISVX] already have a healthy dose of REITs.

- Alec

Not doubting, just curious (thinking about my own allocation) - how do you know this?

 

[Cb]

Btw you can you this spreadsheet to do the simulations to compare results with & without commodities. Also remember this is all past performance - after a point it is a leap of faith because of GIGO.

Simba Backtest

-h

Simba used the S&P Commodity Index for the 1972-2002 period (prior to the inception of the PCRIX, the most frequently recommended Commodity Futures Fund on the DH & ER boards.)

The S&PCI is heavily biased toward energy (~70% IIRC), while the DJ-AIG commodity index that PCRIX contains closer to ~30% energy related commodities.

I used raddr's approach to assemble what I believe is a better synthetic historical representation of PCRIX than the S&PCI...here's a link to the method I used (reviewed by raddr, btw):

http://gnobility.com/Syn_Comm/CCF_1972-2006_san.xls

I use PCRIX actuals for 2003-2006
DJ-AIG plus VIPSX for 2001 -2002
DJ-AIG plus Yahoo's IPS 'category' returns 1997-2001 (prior to VIPSX)
DJ-AIG plus T-Bills for 1991-1996
and the already-collaterallized Chase index for 1972-1990

DJ-AIG plus VIPSX has a very high (.998!) monthly correlation to PCRIX since it's inception so I feel pretty confident about the 2001-2006 era.

The preceding DJ-AIG plus T-Bills is less sound...T-bills are only an OK substitute for TIPS (which didn't exist then), and the pre-1998 DJ_AIG data is somewhat 'datamined' as Rick Ferri frequently points out.

The 1972-1990 Chase index (much more like the DJ-AIG) is perhaps a bit better, using T-Bills & a 'live' index.

I also subtracted the equivalent of PCRIX's substantial 0.74% ER from the various eras as appropriate.

I replaced Simba's CCF dataset in this version of his spreadsheet:

http://gnobility.com/ER/Backtest-Portfolio-returns-My_Syn_CCF.xls

Simba will use that CCF dataset when he updates his spreadsheet with 2007 returns.

hth,

Cb O0

 

[lswswein]

Thanks for that Cb. I liked raddr's data better - great job integrating that

-h

 

[ats5g]

Not doubting, just curious (thinking about my own allocation) - how do you know this?

See REIT Percentages Revealed!

See also Commercial equity real estate: A framework for analysis

- Alec