What am I missing w/this back-tested asset allocation?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sat Dec 10, 2016 9:03 pm

In preparation for annual re-balancing, I used Portfolio Visualizer to back test my current asset allocation (72/28 -- 48% US stocks, 24% foreign stocks, 28% bonds, includes slice/dice allocations to value, REITS, emerging, small cap, foreign bonds, TIPS, etc). My portfolio back-tested pretty well against most standard portfolios when comparing CAGR, IRR, std deviation, Sharpe and Sortino ratios using a $1M starting point and $2000 monthly contributions. The strongest competitor among growth-oriented portfolios seemed to be the Bill Schultheis Coffeehouse Portfolio - it had similar returns over most windows in that timeframe when compared to my portfolio, however with lower volatility.

Before simply shifting to the Schultheis portfolio, I tried improve on it specific to my situation where I wanted to shift to a 60/40 stock/bond asset allocation while still aiming for decent inflation-adjusted growth (I am 54 years old and planning a career downshift with lower income as a transition step towards retirement).

After much tinkering, I was pleasantly surprised that I could come up with a 60/40 asset allocation that seemed to soundly beat every other pre-populated growth-oriented asset allocation model in Portfolio Visualizer in terms of back-tested risk-adjusted return:

My Portfolio - CAGR 9.69% IRR 8.13% Std Deviation 10.75% Sharpe 0.66 Sortino 0.97
Schultheis - CAGR 8.46% IRR 6.80% Std Deviation 9.77% Sharpe 0.58 Sortino 0.83

Here is the portfolio that resulted in the numbers above (40 US Stock, 20 Intl Stock, 40 Bonds):
US Large Cap Value 10.00%
US Small Cap 10.00%
US Small Cap Value 10.00%
REIT 10.00%
Developed Markets 5.00%
Intl Small Cap 5.00%
International Value 5.00%
Emerging Markets 5.00%
Long Term Treasury 10.00%
TIPS 10.00%
Glbl Bond-Unhedged 10.00%
LT Corporate Bonds 10.00%

Since I know that nothing is ever simple or as it appears, I am guessing that I am missing some important points -- before I spend too much time trying to tweak this, what are some of the major issues with this asset allocation that may not be obvious from the numbers? What are the errors in my thinking and logic? Notes: 1) I understand the short-term risks with the possibility of interest increases given the duration of the LT bond portfolios; 2) I tested other shorter intervals with similar positive results
Last edited by boglesmkcents on Sat Dec 10, 2016 9:11 pm, edited 1 time in total.

bayview
Posts: 1878
Joined: Thu Aug 02, 2012 7:05 pm
Location: WNC

Re: What am I missing w/this back-tested asset allocation?

Post by bayview » Sat Dec 10, 2016 9:11 pm

The problem with back-testing is that there is no assurance that history will repeat itself. Unless you have a time machine, you have no idea if you will see the same results.

I pick my (limited number of) funds by what I expect them to do for our investments. Part of that expectation is based on history, of course, but much of it is from an attempt to understand what role the various types of investments can play: total US for growth and inflation protection, Treasuries as a counter-balance when stocks go sour, small-cap international to (one hopes) modify US stock problems, and so forth.

It's definitely fun to play with back-testing, though. :D
The continuous execution of a sound strategy gives you the benefit of the strategy. That's what it's all about. --Rick Ferri

livesoft
Posts: 69586
Joined: Thu Mar 01, 2007 8:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by livesoft » Sat Dec 10, 2016 9:14 pm

I think you have discovered the secret sauce, so I say go for it.

Others have been quite happy with something less spicy. Have you read the Trev H thread: viewtopic.php?t=38374
Wiki This signature message sponsored by sscritic: Learn to fish.

lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 11:55 pm

Re: What am I missing w/this back-tested asset allocation?

Post by lack_ey » Sat Dec 10, 2016 9:17 pm

Nothing looks crazy but it's unclear the extent to which you understand the level of utility of ex-post optimizations over past return data* for constructing asset allocations over future periods.

*which for the record are cobbled from multiple sources, some synthetic like for TIPS, which didn't exist a couple decades ago

What do you think is obvious "from the numbers"? What specifically is your thinking and logic?

What would you estimate the return of this allocation to be the next 10 years and any other properties of interest, and what would you say for an allocation of 40% total US stock, 20% international stock, 40% total bond, for reference?

Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: What am I missing w/this back-tested asset allocation?

Post by Rodc » Sat Dec 10, 2016 9:32 pm

What you are missing is an understanding of something called overfitting

See https://en.wikipedia.org/wiki/Overfitting for a brief introduction.

Google the term to find much more detailed material.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.

stlutz
Posts: 5506
Joined: Fri Jan 02, 2009 1:08 am

Re: What am I missing w/this back-tested asset allocation?

Post by stlutz » Sat Dec 10, 2016 9:43 pm

Try a backtest of backtesting. Go back on portfolio visualizer and see what performed the best for the first half of the period you tested (i.e. 1972-1993). Then see how that did for the second half (1994-2015) relative to a "default" option.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 11, 2016 12:03 am

Thank you all for the very helpful and insightful posts! I do appreciate and understand that past performance is not an indicator of future returns, that efficient frontier graphs change dramatically from one decade/period to the next, the dangers of over-fitting/over-extracting, etc.

My main purpose in choosing many of those investments is that they do tend to be somewhat inversely correlated, and while many of the investments steer towards higher risk/return due to the added risk (small cap, emerging, longer duration bonds, high yield bonds, etc.), the bet I'd be placing is that several of the high volatility investments offset each other in terms of risk, and what is left is the risk premium. It also seems at least A LITTLE safer to bet on long-term bonds now given the recent bond market behavior (although rates admittedly could climb A LOT further).

Another thing that I like with this portfolio (or slight variants of it) is that it seemed to perform well in every 5 and 10 year period that was able to check over the last ~15 years. I recognize this is not a long period of time, however we have had our share of ups and downs during this time. Ant the portfolio seems to cover scenarios where the dollar weakens or stays strong, inflation does or does no rear its head, stocks perform well or the economy tanks, the US markets continue to perform well or strength shifts overseas, etc.

It is definitely not a portfolio built for short-term stability -- the std deviation is fairly high, but as indicated by the Sharpe & Sortino ratios, the results predicted by the last 15+ years show a handsome return for the given risk...of course, past performance is no guarantee of future results, investments carry risks, etc....
Last edited by boglesmkcents on Sun Dec 11, 2016 1:29 am, edited 1 time in total.

Northern Flicker
Posts: 5061
Joined: Fri Apr 10, 2015 12:29 am

Re: What am I missing w/this back-tested asset allocation?

Post by Northern Flicker » Sun Dec 11, 2016 1:19 am

Backtesting with sequential years involves using a highly biased sample, and thus generates poor estimates for expected value or variance of future returns.

User avatar
CaliJim
Posts: 3050
Joined: Sun Feb 28, 2010 8:47 pm
Location: California, near the beach

Re: What am I missing w/this back-tested asset allocation?

Post by CaliJim » Sun Dec 11, 2016 2:41 am

The older I get, the more I appreciate the three fund portfolio.

If your money is in IRAs, great, but if it is taxable, rebalancing to maintain the risk profile can result in an unwanted tax drag.
-calijim- | | For more info, click this Wiki

dcabler
Posts: 1192
Joined: Wed Feb 19, 2014 11:30 am

Re: What am I missing w/this back-tested asset allocation?

Post by dcabler » Sun Dec 11, 2016 7:12 am

Like you, I've done similar backtesting. Heck, you can take the Simba spreadsheet with some simple tweaks and have the excel solver do it's thing and voila! Of course, deciding what to optimize on is the question:
Max CAGR for a particular STDEV
max CAGR for a certain worst case drawdown
max cagr for a certain STDEV in 3 year rolling returns.
The possibilities are, literally, endless.

At the end of the day, this is a fun exercise, but you end up with some portfolios that are on a "knife edge". Meaning that a small change in the % of any one holding might make a large change in the outcome. So, it's always good to do a sensitivity analysis as well.

Also, as you did, you can end up with a highly slice, diced & julienned portfolio that becomes a pain in the rear to rebalance. Then with a little tweaking, you will discover that reducing the number of holdings to a small handful will get you nearly the same performance, is good enough, and can be easily maintained over a lifetime. For many bogleheads here, it's the 3 fund portfolio. Others start there and tilt to SCV. And of course many follow other well known, published lazy portfolios. A number of them appear on Simba's spreadsheet for comparison.

Me:
55/45 Stock/Bond with
Stock: Mid-cap centric leaning towards value, with a dash of small value. Mostly dictated by what's available in my rollover IRA, my current employer's 401K and a deferred compensation account from a previous company and a couple of taxable accounts. Mostly index funds, but some lower cost managed.
Bond: Intermediate Treas index funds and TIPs funds. Sometimes total bond because that's what's available in certain accounts.

There are many roads to Dublin. Some of the roads are harder to drive on than others.....
Last edited by dcabler on Sun Dec 11, 2016 10:43 am, edited 1 time in total.

grok87
Posts: 8917
Joined: Tue Feb 27, 2007 9:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by grok87 » Sun Dec 11, 2016 7:48 am

Listen to bayview and stlutz.

Also backtesting for the past 15 years is basically meaningless. This has been a period of secular decline in interest rates and inflation. To really have a robust back test one would need to go back to say 1970 or so to pick up the stagflation of the late 70s/early 80s.
Or even better to go back to the 1930s to pick up deflation scenarios.

Also LT corporate bonds are not good IMHO.
You might like to read this on the topic...
viewtopic.php?f=10&t=116549
RIP Mr. Bogle.

Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: What am I missing w/this back-tested asset allocation?

Post by Rodc » Sun Dec 11, 2016 8:22 am

boglesmkcents wrote:Thank you all for the very helpful and insightful posts! I do appreciate and understand that past performance is not an indicator of future returns, that efficient frontier graphs change dramatically from one decade/period to the next, the dangers of over-fitting/over-extracting, etc.

Another thing that I like with this portfolio (or slight variants of it) is that it seemed to perform well in every 5 and 10 year period that was able to check over the last ~15 years.
These two quotes indicate you do not have any understanding of what you are doing. Sorry.

There are less than 2 independent data points for 10-year periods in the last 15 years.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.

User avatar
Doc
Posts: 9298
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: What am I missing w/this back-tested asset allocation?

Post by Doc » Sun Dec 11, 2016 8:31 am

Swedroe's black swans book describes a similar portfolio with heavy small value tilt and higher FI to keep risk the same. The idea is to reduce the tails of the distribution while keeping the same mean.

Short book. Easy read. Well worth trip to library or Amazon. Might give you a lot of insight to your idea.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 11, 2016 11:52 am

Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!

User avatar
goingup
Posts: 3717
Joined: Tue Jan 26, 2010 1:02 pm

Re: What am I missing w/this back-tested asset allocation?

Post by goingup » Sun Dec 11, 2016 12:49 pm

A couple observations. I'd want about 70% of that proposed portfolio in tax-advantaged space. Do you have that much room?

I'm mid-50s too, and at this point could not adopt a portfolio like this without incurring lots of capital gains in selling out current positions. Will you have to sell current positions in a taxable account to set up this portfolio?

I think less about portfolio slice and dice optimization and more about taxes, ERs and transaction costs/ :|

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 11, 2016 1:07 pm

goingup wrote:A couple observations. I'd want about 70% of that proposed portfolio in tax-advantaged space. Do you have that much room?

I'm mid-50s too, and at this point could not adopt a portfolio like this without incurring lots of capital gains in selling out current positions. Will you have to sell current positions in a taxable account to set up this portfolio?

I think less about portfolio slice and dice optimization and more about taxes, ERs and transaction costs/ :|
Thank you, goingup -- yes, over 80% of my portfolio is in IRAs and in a 401k, so it is very easy for me to re-balance without incurring any CGs. I am in the same boat with respect to the taxable funds...mostly low basis, so I am waiting for the right time when I retired or in career transition to cash out some of the taxable funds.

User avatar
goingup
Posts: 3717
Joined: Tue Jan 26, 2010 1:02 pm

Re: What am I missing w/this back-tested asset allocation?

Post by goingup » Sun Dec 11, 2016 1:13 pm

boglesmkcents wrote:
goingup wrote:A couple observations. I'd want about 70% of that proposed portfolio in tax-advantaged space. Do you have that much room?

I'm mid-50s too, and at this point could not adopt a portfolio like this without incurring lots of capital gains in selling out current positions. Will you have to sell current positions in a taxable account to set up this portfolio?

I think less about portfolio slice and dice optimization and more about taxes, ERs and transaction costs/ :|
Thank you, goingup -- yes, over 80% of my portfolio is in IRAs and in a 401k, so it is very easy for me to re-balance without incurring any CGs. I am in the same boat with respect to the taxable funds...mostly low basis, so I am waiting for the right time when I retired or in career transition to cash out some of the taxable funds.
That's good that you have all that tax-advantaged space to work with. We're 2/3 taxable so not too nimble with changes.

User avatar
Watty
Posts: 18079
Joined: Wed Oct 10, 2007 3:55 pm

Re: What am I missing w/this back-tested asset allocation?

Post by Watty » Sun Dec 11, 2016 1:38 pm

boglesmkcents wrote:I am guessing that I am missing some important points
1) Taxes and if you are using ETFs you would also have trading commissions and spreads. With that many funds you would be constantly be rebalancing so those costs would add up.

2) Inflation. Your backtesting would be automatically optimizing for a specific past history of inflation. Did you also adjust the $2000 a month contribution for inflation?

3) Interest rates. That would of course favor high interest rate long term bonds as interest rates dropped from double digits to almost zero.

4) You assumed "$1M starting point and $2000 monthly contributions." you would likely get a much result with different assumptions.

User avatar
Taylor Larimore
Advisory Board
Posts: 28952
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: What am I missing w/this back-tested asset allocation?

Post by Taylor Larimore » Sun Dec 11, 2016 1:56 pm

What am I missing w/this back-tested asset allocation?
boglesmkcents:

Back-testing is using past-performance. This is what experts say about using "past performance":
American Association of Individual Investors: "Top Performance lists are dangerous."

Frank Armstrong, financial author: "Rating services such as Morningstar's 'Star Awards' or the 'Forbes Honor Roll' attest to the futility of applying past performance to tomorrow."

Arnott and Bernstein (2002, p. 64): “The investment management industry thrives on the expedient of forecasting the future by extrapolating the past."

Barra Research: "There is no persistence of equity fund performance."

Christine Benz, Morningstars Director of Personal Finance: "When we look at our data, at the factors that are most predictive of good performance going forward, low costs are a much better predictor than is great past performance."

Bruce Berkewitz, manager of the Fairholme Fund (FAIRX), was Morningstar's Manager of the Decade in 2009. In March, 2016 the Fairholme Fund was in the bottom 1% of all funds in its category for 1-year; 3-years; and 5-years.

Wm. Bernstein, author of The Four Pillars of Investing: "For the 20 years from 1970 to 1989, the best performing stock assets were Japanese stocks, U.S. small stocks, and gold stocks. These turned out to be the worst performing assets over the next decade."

Jack Bogle: "The biggest mistake investors make is looking backward at performance and thinking it’ll recur in the future."

Bogleheads' Guide to Investing: "Using past performance to pick tomorrow's winning mutual funds is such a bad idea that the government requires a statement similar to this: "Past performance is no guarantee of future performance." Believe it!"

Jack Brennan, former Vanguard CEO: "Fund ranking is meaningless when based primarily on past performance, as most are."

Burns Advisory tracked the performance of Morningstar's five-star rated stock funds beginning January 1, 1999. Of the 248 stock funds, just four still kept that rank after ten years.

Ben Carlson, author of A Wealth of Common Sense : "Dow Jones looked at nearly 2,900 active mutual funds. Only 2 funds in the top quartile stayed in the top quartile of performance over the next four 1-year periods."

Andrew Clarke, author: "By the time an investment reaches the top of the performance tables, there's a good chance that its run is over. The past is not prologue."

Jonathan Clements, author & former Wall Street Journal columnist: "Suppose you picked stock funds that ranked in their category's top 25% over the past five years. A regular updated study suggests that less than a quarter of these funds will remain in the top 25% over the next five years--even worse than the result you would expect based purely on chance."

Prof. John Cochrane, author: "Past performance has almost no information about future performance."

S.T.Coleridge: "History is a lantern over the stern. It shows where you've been but not where you're going"

Dow Jones Indices Report, June 2015: "The data shows a stronger likelihood for the best-performing funds to become the worst performing funds than vice versa." -- June 2016: "Only 3.7% of large-cap funds maintained top-half performance over five-consecutive 12-month periods. For midcap funds, the comparable figure was 5.79%, and for small-cap funds, it was 7.82%."

Eugene Fama, Nobel Laureate: "Our research on individual mutual funds says that it's impossible to identify true winners on a reliable basis, even if one ignores the costs that active funds impose on investors."

Forbes (2/2/04 issue): "Over the past decade, Morningstar's five-star equity funds have earned an average 5.7% against a 10.3% return for the Wilshire 5000 (Total Stock Market)."

Gensler & Bear, co-authors of The Great Mutual Fund Trap: "Of the fifty top-performing funds in 2000, not a single one appeared on the list in either 1999 or 1998."

Ken Hebner's CGM Focus Fund was the top U.S. equity fund in 2007. In November 2009, it ranked in the bottom 1% of its category.

Mark Hulbert (12-31-2014): "Consider a hypothetical portfolio that each year followed the investment newsletter portfolio that, among the more than 500 tracked by The Hulbert Financial Digest, had the best record during the previous calendar year. Over the past 20 years, that portfolio would have been a disaster, producing an annualized loss of more than -17%."

JPMorgan Chase claimed that 97% of their alternate-asset mutual funds beat their benchmark during the 10-year period ending December, 2013. Morningstar reported that only 33% beat their benchmark during the same period (past-performance calculations differ).

Arthur Levitt, SEC Commissioner: "A mutual fund's past performance, which is the first feature that investors consider when choosing a fund, doesn't predict future performance."

Peter Lynch's Fidelity Magellan Fund (FMAGX), once the worlds largest and most successful mutual fund, is now in the bottom 10% of its category (Sept 7-2016) for 10-year return

Burton Malkiel, author of the classic Random Walk Down Wall Street: "I have examined the lack of persistency in fund returns over periods from the 1960s through the early 2000s.--There is no persistency to good performance. It is as random as the market."

Mercer Investment Consulting from a study of over 12,000 institutional managers: "Excellent recent performance not only doesn't guarantee future results but generally leads to under-performance in the subsequent period."

Bill Miller, former manager of Legg Mason Value Trust (LMVTX), was the only manager to outperform The S&P 500 Index for 15 consecutive years. On 9/7/2016 Miller’s fund is in the bottom 1% for 15 year returns.

Mark Miller, financial author and journalist: "Only 7.33% of domestic equity funds that were in the top quartile of performance in March 2014 were still there two years later."

Morningstar: "The star rating was not designed to have predictive ability about future performance."

Ron Ross, author of The Unbeatable Market: "Extensive studies by Davis, Brown & Groetzman, Ibbotson, Elton et al, all confirmed there is no significant persistance in mutual fund performance."

Bill Schultheis, adviser and author of The Coffeehouse Investor: "Using past performance numbers as a method for choosing mutual funds is such a lousy idea that mutual fund companies are required by law to tell you it is a lousy idea."

Standard & Poor's Persistence Scorecard (Dec-2014): "The data show a stronger likelihood for the best-performing funds to become the worst-performing funds than vice versa. Of 421 funds that were in the bottom quartile, 14.45% moved to the top quartile over the five year horizon, while 27.08% of the 421 funds that were in the top quartile moved into the bottom quartile during the same period."

Larry Swedroe, author of many finance books: "The 44 Wall Street Fund was the top performing fund over the decade of the 1970s. It ranked as the single worst performing fund of the 1980's losing 73%. -- If you are going to use past performance to predict the future winners, the evidence is strong that your approach is highly likely to fail."

David Swensen, Yale's Chief Investment Officer: "Chasing performance is the biggest mistake investors make. If anything, it is a perverse indicator."

Tweddell & Pierce, co-authors of Winning With Mutual Funds: "Numerous studies have shown that using superior past performance is no better than random selection."

Eric Tyson, author of Mutual Funds for Dummies (2010 edition): "Of the number one top-performing stock and bond funds in each of the last 20 years, a whopping 80% of them subsequently performed worse than the average fund in their peer group over the next 5 to 10 years! Some of these former #1 funds actually went on to become the worst-performing funds in their particular category."

Value Line selected Garret Van Wagoner "Mutual fund Manager of the Year" in 1999. In August 2009, Van Wagoner's Emerging Growth Fund was the worst performing U.S. stock fund over the past 10 years.

Vanguard Study: "Persistence of performance among past winners is no more predictable than a flip of a coin."

Jason Zweig, author and Wall Street Journal columnist: "Buying funds based purely on their past performance is one of the stupidest things an investor can do."
Read the Simplicity link below.

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

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 11, 2016 2:09 pm

Taylor Larimore wrote:
What am I missing w/this back-tested asset allocation?
boglesmkcents:

Back-testing is using past-performance. This is what experts say about using "past performance":

Thank you Taylor! Excellent reminders to take the data from back testing with a grain of salt. Aside, from the excellent quotes you provided, the periodic table of investments and the varied efficient frontier graphs are also good reminders.

All this said, I believe that there is value in using the back-test data as input, especially as it relates to Sharpe ratio and volatility, to check different time periods, and to check assumptions about correlation in those varied economic climates...does that make sense?

User avatar
Taylor Larimore
Advisory Board
Posts: 28952
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: What am I missing ?

Post by Taylor Larimore » Sun Dec 11, 2016 6:01 pm

I believe that there is value in using the back-test data as input, especially as it relates to Sharpe ratio and volatility, to check different time periods, and to check assumptions about correlation in those varied economic climates...does that make sense?
boglesmkcents:

After spending many years trying to beat-the-market, I learned to invest successfully by listening to independent experts who know more than I do (not the self-serving financial industry) .

If I could start over today, I would invest entirely in The Three-Fund Portfolio. You can read its many advantages here:

The Three-Fund Portfolio

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

grok87
Posts: 8917
Joined: Tue Feb 27, 2007 9:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by grok87 » Sun Dec 11, 2016 7:40 pm

boglesmkcents wrote:Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!
Hi boglesmkcents,
here is some more data for you
http://mba.tuck.dartmouth.edu/pages/fac ... brary.html
in particular, small cap stocks and small cap value stocks got crushed in the great depression.
RIP Mr. Bogle.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing ?

Post by boglesmkcents » Sun Dec 11, 2016 10:38 pm

Taylor Larimore wrote:
I believe that there is value in using the back-test data as input, especially as it relates to Sharpe ratio and volatility, to check different time periods, and to check assumptions about correlation in those varied economic climates...does that make sense?
boglesmkcents:

After spending many years trying to beat-the-market, I learned to invest successfully by listening to independent experts who know more than I do (not the self-serving financial industry) .

If I could start over today, I would invest entirely in The Three-Fund Portfolio. You can read its many advantages here:

The Three-Fund Portfolio

Best wishes.
Taylor
Thank you Taylor -- I definitely have a struggle going on in my mind between the simple (yet powerful) 3 or 4 fund portfolio, vs the modestly more complex 5-10 fund (moderate) slice/dice approach...I will definitely re-read the post you reference above!

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 11, 2016 10:45 pm

grok87 wrote:
boglesmkcents wrote:Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!
Hi boglesmkcents,
here is some more data for you
http://mba.tuck.dartmouth.edu/pages/fac ... brary.html
in particular, small cap stocks and small cap value stocks got crushed in the great depression.
Thank you, grok87 -- important food for thought re: what could happen with a portfolio so heavily weighted to small caps...yes, I see the data now in the Simba spreadsheet for the small caps in the late 20's and 30's - yikes! I also really appreciate the pointer to that treasure trove of Dartmouth data! Lots to consider, for sure!...

grok87
Posts: 8917
Joined: Tue Feb 27, 2007 9:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by grok87 » Sun Dec 11, 2016 10:49 pm

boglesmkcents wrote:
grok87 wrote:
boglesmkcents wrote:Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!
Hi boglesmkcents,
here is some more data for you
http://mba.tuck.dartmouth.edu/pages/fac ... brary.html
in particular, small cap stocks and small cap value stocks got crushed in the great depression.
Thank you, grok87 -- important food for thought re: what could happen with a portfolio so heavily weighted to small caps...yes, I see the data now in the Simba spreadsheet for the small caps in the late 20's and 30's - yikes! I also really appreciate the pointer to that treasure trove of Dartmouth data! Lots to consider, for sure!...
you're welcome! have fun...
RIP Mr. Bogle.

carolinaman
Posts: 3939
Joined: Wed Dec 28, 2011 9:56 am
Location: North Carolina

Re: What am I missing w/this back-tested asset allocation?

Post by carolinaman » Mon Dec 12, 2016 8:02 am

The older I get, the more I appreciate a simple 3 fund portfolio. To each his own, but this portfolio seems far to complex and granular for me. I have always wondered why a 5% or 10% position in anything would have a material affect on your total portfolio.

Others have warned of the dangers of backtesting, especially of a period as limited as 15 years, and I echo their sentiments. Will the next 5, 10 or 15 years resemble the last 15 years. Who knows but my best is it will be different.

dcabler
Posts: 1192
Joined: Wed Feb 19, 2014 11:30 am

Re: What am I missing w/this back-tested asset allocation?

Post by dcabler » Mon Dec 12, 2016 10:05 am

grok87 wrote:
boglesmkcents wrote:
grok87 wrote:
boglesmkcents wrote:Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!
Hi boglesmkcents,
here is some more data for you
http://mba.tuck.dartmouth.edu/pages/fac ... brary.html
in particular, small cap stocks and small cap value stocks got crushed in the great depression.
Thank you, grok87 -- important food for thought re: what could happen with a portfolio so heavily weighted to small caps...yes, I see the data now in the Simba spreadsheet for the small caps in the late 20's and 30's - yikes! I also really appreciate the pointer to that treasure trove of Dartmouth data! Lots to consider, for sure!...
you're welcome! have fun...
Also note that the Simba spreadsheet is a work in progress. If you go to the thread where it's discussed, you'll see that there is also an unreleased, working copy where siamond is adding new info (including some series that do go back quite far) and there are always good discussions about where the best sources of data are or the best way to create a synthetic return series for indices that didn't exist. For me, I do like backtesting, but only for general trends. It's more of a sledgehammer than a scalpel.

As for small caps during the depression, I seem to recall a statement somewhere about the number of stocks that were actually included in the fama-french data or even the SIMBA pseudo-CRSP data. My recollection is that the number of stocks considered small cap for that period was quite small - very high granularity compared to today. Yet another reason to not overly fine-tune an AA based on backtesting.

staythecourse
Posts: 6993
Joined: Mon Jan 03, 2011 9:40 am

Re: What am I missing w/this back-tested asset allocation?

Post by staythecourse » Mon Dec 12, 2016 11:04 am

The only thing I can gaurantee is that your portfolio will NOT be the best performer going forward. There is only ONE portfolio on the EF and it will only be known in retrospect.

But then again who cares. The point is NOT to aim for the best but to be good enough. That means one that you will stick with as staying the course is more important then whatever backtesting proves to be right or wrong.

Good luck.
"The stock market [fluctuation], therefore, is noise. A giant distraction from the business of investing.” | -Jack Bogle

Gropes & Ray
Posts: 1068
Joined: Wed Jul 16, 2014 7:28 am

Re: What am I missing w/this back-tested asset allocation?

Post by Gropes & Ray » Mon Dec 12, 2016 11:20 am

I am exhausted by the thought of rebalancing that portfolio. :beer

Have you looked into the Larry Portfolio? I haven't looked at it extensively, but I think it uses small caps to try to get big gains and then a large bond holding to even out the swings.

dcabler
Posts: 1192
Joined: Wed Feb 19, 2014 11:30 am

Re: What am I missing w/this back-tested asset allocation?

Post by dcabler » Mon Dec 12, 2016 5:14 pm

grok87 wrote:
boglesmkcents wrote:Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!
Hi boglesmkcents,
here is some more data for you
http://mba.tuck.dartmouth.edu/pages/fac ... brary.html
in particular, small cap stocks and small cap value stocks got crushed in the great depression.
A couple of articles I ran across regarding the CRSP dataset, especially with regards to small cap stocks back then

- the original paper: https://www.researchgate.net/publicatio ... _US_Stocks

- Excerpts from the paper by another author focused on small cap: http://theirrelevantinvestor.com/2015/1 ... -told-you/

grok87
Posts: 8917
Joined: Tue Feb 27, 2007 9:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by grok87 » Mon Dec 12, 2016 8:22 pm

dcabler wrote:
grok87 wrote:
boglesmkcents wrote:Thank you everyone for the really excellent, direct and candid advise as always! This gives me very helpful perspective, and definitely some areas to research.

I appreciate the importance of testing back through longer periods, however many of the asset classes did not have the long history, and I had not looked into SIMBA's spreadsheet (assumed it was the same data as Portfolio Visualizer). I just downloaded SIMBA's spreadsheet -- wow, what an amazing tool! The only downside is that now I will not see my family for weeks. Portfolio Visualizer is also great,but it is nice to have at least synthetic views of more asset classes deeper in time.

I will likely have many more questions, but for now, I have plenty to keep me busy. Many thanks again for all the feedback and guidance!
Hi boglesmkcents,
here is some more data for you
http://mba.tuck.dartmouth.edu/pages/fac ... brary.html
in particular, small cap stocks and small cap value stocks got crushed in the great depression.
A couple of articles I ran across regarding the CRSP dataset, especially with regards to small cap stocks back then

- the original paper: https://www.researchgate.net/publicatio ... _US_Stocks

- Excerpts from the paper by another author focused on small cap: http://theirrelevantinvestor.com/2015/1 ... -told-you/
thanks for the link
so I guess the argument is that the small cap premium before 1970 may be overstated. Doesn't seem to necessarily argue against the idea that small caps did worse than average in the depression/deflation years. one explanation i have heard is that small caps tend to be more leveraged and hence get hurt more badly in deflationary scenarios...
RIP Mr. Bogle.

User avatar
Portfolio7
Posts: 701
Joined: Tue Aug 02, 2016 3:53 am

Re: What am I missing w/this back-tested asset allocation?

Post by Portfolio7 » Tue Dec 13, 2016 5:31 am

I only like LT bonds in conjunction with very high equity portfolios. We've had a 30 year bull in bonds, which makes LT bonds look really good in even 40 year backtests. There are no guarantees going forward, but in a 60/40 portfolio Intermediate Treasuries have historically given a lot of bang for the buck - or a Total Bond Fund is very similar in performance and more diversified. There is some discussion of this in some boglehead discussions, and also some of the older posts at 'A wealth of common sense'.
"An investment in knowledge pays the best interest" - Benjamin Franklin

Valuethinker
Posts: 39235
Joined: Fri May 11, 2007 11:07 am

Re: What am I missing w/this back-tested asset allocation?

Post by Valuethinker » Tue Dec 13, 2016 5:41 am

Portfolio7 wrote:I only like LT bonds in conjunction with very high equity portfolios. We've had a 30 year bull in bonds, which makes LT bonds look really good in even 40 year backtests. There are no guarantees going forward, but in a 60/40 portfolio Intermediate Treasuries have historically given a lot of bang for the buck - or a Total Bond Fund is very similar in performance and more diversified. There is some discussion of this in some boglehead discussions, and also some of the older posts at 'A wealth of common sense'.
I absolutely agree with what you say, and an IT Treasury fund is very appropriate for this. TBM is less so, but the actual difference in practice is small enough that if that's what an investor chooses, it is unlikely to be a very harmful choice.

Just on LT Corporate Bonds

LT Corporate bonds (which I think the OP referenced as an investment) are NOT a good idea:

- credit risk rises over time. Think of the supercompanies of 40 years ago: Polaroid, Xerox, Kodak. GM etc. You've got no protection against a CEO who comes in, buys back a lot of shares by issuing bonds, then eventually the company goes Chapter 11

- sectoral risk is not diversified fully - different sectors go through cycles of issuing bonds eg telecoms up to the 2001 dot com crash (Worldcom!)

- many of the bonds contain call provisions which means the bonds are redeemed just when the investor doesn't want them to be

For similar reasons, LT Agency securities (GNMA FNMA FMAC) are also not a good idea-- no credit risk, by prepayment risk.

You have to have LT US Treasury bonds.

grok87
Posts: 8917
Joined: Tue Feb 27, 2007 9:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by grok87 » Tue Dec 13, 2016 6:36 am

Valuethinker wrote:
Portfolio7 wrote:I only like LT bonds in conjunction with very high equity portfolios. We've had a 30 year bull in bonds, which makes LT bonds look really good in even 40 year backtests. There are no guarantees going forward, but in a 60/40 portfolio Intermediate Treasuries have historically given a lot of bang for the buck - or a Total Bond Fund is very similar in performance and more diversified. There is some discussion of this in some boglehead discussions, and also some of the older posts at 'A wealth of common sense'.
I absolutely agree with what you say, and an IT Treasury fund is very appropriate for this. TBM is less so, but the actual difference in practice is small enough that if that's what an investor chooses, it is unlikely to be a very harmful choice.

Just on LT Corporate Bonds

LT Corporate bonds (which I think the OP referenced as an investment) are NOT a good idea:

- credit risk rises over time. Think of the supercompanies of 40 years ago: Polaroid, Xerox, Kodak. GM etc. You've got no protection against a CEO who comes in, buys back a lot of shares by issuing bonds, then eventually the company goes Chapter 11

- sectoral risk is not diversified fully - different sectors go through cycles of issuing bonds eg telecoms up to the 2001 dot com crash (Worldcom!)

- many of the bonds contain call provisions which means the bonds are redeemed just when the investor doesn't want them to be

For similar reasons, LT Agency securities (GNMA FNMA FMAC) are also not a good idea-- no credit risk, by prepayment risk.

You have to have LT US Treasury bonds.
+1
RIP Mr. Bogle.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 5:40 pm

I am back with an update after taking a week to digest all the great feedback I received, and to adapt my approach accordingly.

I have not yet had time to read all the suggested material referenced in the suggestions, but I did incorporate much of the advice I received, including:
• Avoid corporate bonds in favor of Treasures
• OK to use back-testing as a tool, but more importantly consider the role of each fund
• Utilize Simba’s spreadsheet to back-test over a longer period (now to 1972, and spot check back to 1927)
• Consider different time periods (for example 1972-1993 and 1993 – 2015, 1920s/1930s)
• Don’t bother with slices as small as 5% due to limited impact vs. the effort involved to re-balance
• Incorporate ideas from Larry Swedroe writings/portfolios re: Black Swans/Fat Tails
• Review related slice/dice Trev H thread

After considering all this and doing further analysis, I have adjusted the 11 fund portfolio down to the following 5 fund portfolio:
40% US Equities:
o 30% SCV
o 10% REIT
20% Intl Equities
o 20% Intl Small Cap
40% Bonds:
o 30% LTT
o 10% TIPS

As mentioned earlier, this is for our retirement portfolio (55 years old, investing for long-term, most in tax-deferred accts). Basic idea is to capture higher risk premium using riskier yet somewhat negatively correlated funds and a heavy dose of both aggressive stock funds and longer-term bonds to maximize return and yet minimize volatility (taking Larry Swedroe's ideas a step further).

This portfolio seems to perform very well vs. many of the coffee house portfolios for the 1972-2015 time frame (often 1-2% higher returns with lower volatility, back-tested 44 years). The worst relative time period is ~1980-1998, but it holds its own there as well. For the full 44 year window 1972-2015, the CAGR approaches 12% w/std deviation a respectable 11.8%. Our current portfolio is 48/24/28 US/Intl/Bonds – new this 40/20/40 mix increases historic returns while reducing volatility. In practice, I’ll probably use the portfolio as a “concept portfolio”, and will probably moderate the aggressive portfolio above with some ITT, TSM and ISM (a moderated approach would have performed better in the depression years, and in the 1980-1998 time-frame).

I understand that past performance cannot be relied on for future performance…I am only trying to understand if a portfolio like this represents a good risk/return trade-off and as the subject of my original post states, what possible pitfalls am I still be missing after trying to address the more obvious short-comings from my initial post?

I learned a lot from all the suggestions (thank you!) and now I am very interested in any thoughts about implementing this updated approach!

livesoft
Posts: 69586
Joined: Thu Mar 01, 2007 8:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by livesoft » Sun Dec 18, 2016 6:03 pm

boglesmkcents wrote:[…]
• Review related slice/dice Trev H thread

[...'
In practice, I’ll probably use the portfolio as a “concept portfolio”, and will probably moderate the aggressive portfolio above with some ITT, TSM and ISM (a moderated approach would have performed better in the depression years, and in the 1980-1998 time-frame).
If you take your concept portfolio and add in Total US Stock Market and International Stock Market, isn't that pretty much a slice/dice like Trev H showed?
Wiki This signature message sponsored by sscritic: Learn to fish.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 6:08 pm

livesoft wrote:
boglesmkcents wrote:[…]
• Review related slice/dice Trev H thread

[...'
In practice, I’ll probably use the portfolio as a “concept portfolio”, and will probably moderate the aggressive portfolio above with some ITT, TSM and ISM (a moderated approach would have performed better in the depression years, and in the 1980-1998 time-frame).
If you take your concept portfolio and add in Total US Stock Market and International Stock Market, isn't that pretty much a slice/dice like Trev H showed?
Yes, I think so -- however I am really asking about my 5 fund portfolio with equities concentrated in SCV (30%) and international small cap (20%, and bonds concentrated in LTT (30%), plus some TIPS (10%) and REITS (10%) to add a little diversification.

livesoft
Posts: 69586
Joined: Thu Mar 01, 2007 8:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by livesoft » Sun Dec 18, 2016 6:15 pm

The pitfalls you are missing are the normal behavioral and emotional ones. Can you perform like a computer algorithm? Can you rebalance like madsinger did: viewtopic.php?p=3143586#p3143586 ? He bought bonds a few days before the FOMC meeting announcement instead of waiting. He was too much of a robot.
Wiki This signature message sponsored by sscritic: Learn to fish.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 6:25 pm

livesoft wrote:The pitfalls you are missing are the normal behavioral and emotional ones. Can you perform like a computer algorithm? Can you rebalance like madsinger did: viewtopic.php?p=3143586#p3143586 ? He bought bonds a few days before the FOMC meeting announcement instead of waiting. He was too much of a robot.
Thanks Livesoft -- I've considered re-balancing based on bands, and I've considered time intervals from monthly to every 2 years. So far, for the past few years, I've stayed the course re-balancing annually during the last week of December. I have been pretty good about not letting my emotions getting in the way -- I have learned that most of the time when I think I know what markets will do next, I am wrong (>> 50% of the time).

Quark
Posts: 1045
Joined: Sun Nov 01, 2015 5:32 pm

Re: What am I missing w/this back-tested asset allocation?

Post by Quark » Sun Dec 18, 2016 6:29 pm

Vanguard has a research paper on forecasting stock returns at https://personal.vanguard.com/pdf/s338.pdf

Two major lessons:

1) Historical returns are not the best method for forecasting the future.

2) The best method doesn't work very well, explaining only about 40% of the time variation in net-of-inflation returns.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 6:51 pm

Quark wrote:Vanguard has a research paper on forecasting stock returns at https://personal.vanguard.com/pdf/s338.pdf

Two major lessons:

1) Historical returns are not the best method for forecasting the future.

2) The best method doesn't work very well, explaining only about 40% of the time variation in net-of-inflation returns.
Thanks, Quark -- I strongly agree with these important lessons. This is why I have the portfolio loaded up with index investments that are inversely correlated, because I know that nobody can predict the future.

It does seem that this portfolio has done well in many different market climates, and outperforms simple coffee house portfolios substantially on a risk-adjusted basis over the long-term, and in market climates where it under-performs, it seems to do so modestly.

Since the data seems to indicate that this portfolio has out-performed many traditional mixes over 44 years, I am thinking it is less risky than a more vanilla portfolio such as 40% TSM/20% TISM/40% TBM even though the proposed portfolio is made up of a few very volatile components.

In my mind, long-term risk does not equal short-term volatility for a long-term portfolio, and in this case, the volatility seems to cancel out and the extra return showing in the strong long-term performance is the reward for accepting risk in the components.

That is my hypothesis that I am testing here. Eager to hear who agrees/disagrees, and why?

Quark
Posts: 1045
Joined: Sun Nov 01, 2015 5:32 pm

Re: What am I missing w/this back-tested asset allocation?

Post by Quark » Sun Dec 18, 2016 7:00 pm

boglesmkcents wrote:
Quark wrote:Vanguard has a research paper on forecasting stock returns at https://personal.vanguard.com/pdf/s338.pdf

Two major lessons:

1) Historical returns are not the best method for forecasting the future.

2) The best method doesn't work very well, explaining only about 40% of the time variation in net-of-inflation returns.
Thanks, Quark -- I strongly agree with these important lessons. This is why I have the portfolio loaded up with index investments that are inversely correlated, because I know that nobody can predict the future.
Your tense is off. Investments have been inversely correlated, not are inversely correlated. What matters for investing is whether they will be inversely correlated.

What have you found that is inversely correlated? Weak correlation is easy to find, but true inverse correlation is rare, especially for index funds.
boglesmkcents wrote:It does seem that this portfolio has done well in many different market climates, and outperforms simple coffee house portfolios substantially on a risk-adjusted basis over the long-term, and in market climate where it under-performs, it seems to do so modestly.
That's just forecasting based on historical returns, which we seem to agree doesn't work very well.
boglesmkcents wrote:Since the data seems to indicate that this portfolio has out-performed many traditional mixes over 44 years, I am thinking it is less risky than a more vanilla portfolio such as 40% TSM/20% TISM/40% TBM even though it is made up of a few very volatile components.
It's very easy to find portfolios that have outperformed. It's important to realize how little that means for the future.
boglesmkcents wrote:In my mind, risk does not equal volatility, and in this case, the volatility seems to cancel out and what is left is reward for accepting risk in the components, even though the risk has been cancelled out.
First you write that risk does not equal volatility, then you write that canceling out volatility cancels risk.

I'd recommend re-reading the comments in this thread. As Rodc wrote, "These ... quotes indicate you do not have any understanding of what you are doing. Sorry."

grok87
Posts: 8917
Joined: Tue Feb 27, 2007 9:00 pm

Re: What am I missing w/this back-tested asset allocation?

Post by grok87 » Sun Dec 18, 2016 7:02 pm

boglesmkcents wrote:
After considering all this and doing further analysis, I have adjusted the 11 fund portfolio down to the following 5 fund portfolio:
40% US Equities:
o 30% SCV
o 10% REIT
20% Intl Equities
o 20% Intl Small Cap
40% Bonds:
o 30% LTT
o 10% TIPS
i think this looks good. Yes if you are going to heavily tilt to small caps you want most of your bonds in long term treasuries.
THe LTT may get killed in a high inflation scenario, but hopefully your small caps will do well then
RIP Mr. Bogle.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 7:22 pm

Quark wrote: Your tense is off. Investments have been inversely correlated, not are inversely correlated. What matters for investing is whether they will be inversely correlated.

What have you found that is inversely correlated? Weak correlation is easy to find, but true inverse correlation is rare, especially for index funds.
Fair point...I agree with your clarification
boglesmkcents wrote:It does seem that this portfolio has done well in many different market climates, and outperforms simple coffee house portfolios substantially on a risk-adjusted basis over the long-term, and in market climate where it under-performs, it seems to do so modestly.
Quark wrote: That's just forecasting based on historical returns, which we seem to agree doesn't work very well.
To me, it makes sense to use the data as an important consideration...also important to realize it far from guarantees future performance.

boglesmkcents wrote:Since the data seems to indicate that this portfolio has out-performed many traditional mixes over 44 years, I am thinking it is less risky than a more vanilla portfolio such as 40% TSM/20% TISM/40% TBM even though it is made up of a few very volatile components.
Quark wrote: It's very easy to find portfolios that have outperformed. It's important to realize how little that means for the future.
Well, to me, 44 years worth of data is important to consider.
boglesmkcents wrote:In my mind, risk does not equal volatility, and in this case, the volatility seems to cancel out and what is left is reward for accepting risk in the components, even though the risk has been cancelled out.
Quark wrote: First you write that risk does not equal volatility, then you write that canceling out volatility cancels risk.
What I am trying to say is that there is a reward for accepting the volatility (referred to by many as the risk premium), and if you accept the volatility/risk, there is a reward when the investment(s) is/are held for the long term.
Quark wrote: I'd recommend re-reading the comments in this thread. As Rodc wrote, "These ... quotes indicate you do not have any understanding of what you are doing. Sorry."
Well, I am trying to learn from these comments...I tried to address the shortcoming that Rodc pointed out, but apparently my logic is still mis-guided.

I accept that I am probably wrong and do not have any understanding of what I am doing, but what I am currently trying to understand is why one would prefer to hold investment portfolios that 44 years of data indicate have under-performed (from a risk-return standpoint) vs. another portfolio that has historically, and through different economic climates, had higher returns with lower volatility.

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 7:31 pm

grok87 wrote:
boglesmkcents wrote:
After considering all this and doing further analysis, I have adjusted the 11 fund portfolio down to the following 5 fund portfolio:
40% US Equities:
o 30% SCV
o 10% REIT
20% Intl Equities
o 20% Intl Small Cap
40% Bonds:
o 30% LTT
o 10% TIPS
i think this looks good. Yes if you are going to heavily tilt to small caps you want most of your bonds in long term treasuries.
THe LTT may get killed in a high inflation scenario, but hopefully your small caps will do well then
Thank you, grok87 -- your constructive comments have been very helpful in showing me how to improve the portfolio -- I can see where my earlier version that had LT Corp bonds might not have functioned well in a very poor economy. I also have TIPS in there to help in a high inflation environment, and REITS may help here as well.

User avatar
Tyler9000
Posts: 479
Joined: Fri Aug 21, 2015 11:57 am

Re: What am I missing w/this back-tested asset allocation?

Post by Tyler9000 » Sun Dec 18, 2016 7:39 pm

stlutz wrote:Try a backtest of backtesting. Go back on portfolio visualizer and see what performed the best for the first half of the period you tested (i.e. 1972-1993). Then see how that did for the second half (1994-2015) relative to a "default" option.
+1

Also, try this: https://portfoliocharts.com/portfolio/heat-map/

You can plug in your AA and it will calculate every investing period since 1972 simultaneously (990 total). That will give you a nice perspective for the full set of outcomes for a specific portfolio regardless of start date. Then compare that chart to those for other portfolios.

Just looking at a single backtested number can be quite deceptive, but you can learn a lot about a portfolio by studying how it has performed in all types of economic conditions.

heyyou
Posts: 3595
Joined: Tue Feb 20, 2007 4:58 pm

Re: What am I missing w/this back-tested asset allocation?

Post by heyyou » Sun Dec 18, 2016 7:50 pm

What am I missing w/this back-tested asset allocation?
You asked the question, but I do apologize for the tone of my response.

Paraphrasing others, you will not get the long term average return, you will get the specific return for your dollar weighted period. Consider how often your research has found a new favorite. That will not entirely cease, even after you stop looking. There will always be a better, recent performing mix publicized than whatever you chose a few years ago. The Callan Periodic Table also shows loss numbers for how poorly some of those higher performing sub-asset classes can do, just after a few good years.

The mostly high risk equities with 2/3 bonds portfolios are held by very experienced investors with enough wealth in the bonds that growth of the equity portion doesn't matter to their lifestyle, but you seem to be adopting double their equity risk with your allocation because you want the gains shown in the back testing. If it was that easy, the hedge funds and the actively managed mutual funds would be there ahead of you. Why are they not doing that?

Michael McClung in Living Off Your Money found that somewhat equal weighted portfolios were better over retirement length periods than more concentrated ones, but do what suits you. I agree that there are numerous portfolios that will perform well enough.

The suggestion to check other periods is good too. Large Caps with plenty of Large Growth companies did quite well for a long time, late in the last century. Those were high paying work years with decades of savings for many of the currently retired posters here, but long term, Large Growth's volatility is not as good during retirement. Be aware of how high volatility can be a problem when you are making periodic withdrawals.

An alternative to seeking the optimal portfolio is to just save more than others and you will then retire with more assets. As Jack Bogle said, "It is simple, but not easy."

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 7:55 pm

Tyler9000 wrote:
stlutz wrote:Try a backtest of backtesting. Go back on portfolio visualizer and see what performed the best for the first half of the period you tested (i.e. 1972-1993). Then see how that did for the second half (1994-2015) relative to a "default" option.
+1

Also, try this: https://portfoliocharts.com/portfolio/heat-map/

You can plug in your AA and it will calculate every investing period since 1972 simultaneously (990 total). That will give you a nice perspective for the full set of outcomes for a specific portfolio regardless of start date. Then compare that chart to those for other portfolios.

Just looking at a single backtested number can be quite deceptive, but you can learn a lot about a portfolio by studying how it has performed in all types of economic conditions.
Thank you, Tyler9000 -- I will definitely check this out!

Topic Author
boglesmkcents
Posts: 88
Joined: Tue Jul 24, 2012 4:57 pm

Re: What am I missing w/this back-tested asset allocation?

Post by boglesmkcents » Sun Dec 18, 2016 8:26 pm

heyyou wrote:
What am I missing w/this back-tested asset allocation?
You asked the question, but I do apologize for the tone of my response.
No need to apologize -- I found your response very constructive!
heyyou wrote: There will always be a better, recent performing mix publicized than whatever you chose a few years ago. The Callan Periodic Table also shows loss numbers for how poorly some of those higher performing sub-asset classes can do, just after a few good years.

The mostly high risk equities with 2/3 bonds portfolios are held by very experienced investors with enough wealth in the bonds that growth of the equity portion doesn't matter to their lifestyle, but you seem to be adopting double their equity risk with your allocation because you want the gains shown in the back testing. If it was that easy, the hedge funds and the actively managed mutual funds would be there ahead of you. Why are they not doing that?
I am not sure -- perhaps some hedge funds and actively managed funds are using some of these approaches? Not saying they are, I don't know...
heyyou wrote: Michael McClung in Living Off Your Money found that somewhat equal weighted portfolios were better over retirement length periods than more concentrated ones, but do what suits you. I agree that there are numerous portfolios that will perform well enough.
I had already backed off from 72% equities to 60% -- the 72% has served me well, but we are getting to an age where a downshift seems appropriate.
heyyou wrote: The suggestion to check other periods is good too. Large Caps with plenty of Large Growth companies did quite well for a long time, late in the last century. Those were high paying work years with decades of savings for many of the currently retired posters here, but long term, Large Growth's volatility is not as good during retirement. Be aware of how high volatility can be a problem when you are making periodic withdrawals.
Understood -- what I am most looking for is any feedback on whether the combination of domestic/international small caps combined with LTT's/TIPS/REITS have a high likelihood (not guarantee!) to be an combination in a portfolio based on historical patterns and based on the logic that they are likely move in opposite directions and that all (especially small caps and LTTs) carry a significant risk premium.
heyyou wrote: An alternative to seeking the optimal portfolio is to just save more than others and you will then retire with more assets. As Jack Bogle said, "It is simple, but not easy."
I have been a buy and hold, index investor with very generic asset allocations for 30 years, and we have always saved ~20% of our income. The model has served us well, and I have tremendous respect for Jack Bogle, the Bogleheads, and the general concepts so strongly advocated here. So bottom line is that I have always done what most people here are telling me to do. That said, I strongly appreciate the difference made if once can increase return by even .5% per year, without increasing risk appreciably (or at all), so basically I am trying to optimize future returns and capital preservation for our hard earned and hard saved assets.

Quark
Posts: 1045
Joined: Sun Nov 01, 2015 5:32 pm

Re: What am I missing w/this back-tested asset allocation?

Post by Quark » Sun Dec 18, 2016 8:45 pm


Post Reply