10 Year Retrospective

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

10 Year Retrospective

Post by Robert T »

.
As its now been 10 years since setting up my portfolio - a quick retrospective (FWIW).

Investment Policy (unchanged for last 10 years)

Image

Expected long-term
Annualized Return = 7.5%
Downside (% 1-yr loss) = 30-35%

Actual over last ten years
Annualized Return 2003-12 = 10.1%
Largest one year loss (2008) = 28.7%

Estimated source of return 2003-12

+1.7% = T-bills
+5.0% = Equity (from 75% equity exposure)
+0.9% = Term (from use of intermediate [3-7 yr] treasuries, with 0.5 term load target)
+1.3% = Intl diversification (from EAFE and EM exposure)
+1.2% = Value + size tilt (from 0.4/0.2 value/size load target)
+10.1% = Total Return

Over the last 10 years everything worked, in that the equity, term, value, and size tilt contributions were all positive, and intl. diversification added additional returns (particularly exposure to EM) – more so than expected with actual returns of 10.1% compared to a long-term expectation of 7.5%. The 2008 decline was close to the range of expected one year losses, and was good to know it was broadly in-line with my tolerable loss! I plan to keep the same investment policy going forward.

Some general observations over the last 10 years

Human capital returns (and associated savings) matter: For early investors, income (and associated savings) can be a large share of investments and investment returns. Overtime this declines and eventually reverses. Percent income gains can sometimes be larger than investment returns – and its important to try to ensure these income gains translate into higher savings. Its often very easy to expand into the space (income) we have (bigger house, better car, iphone, etc etc, etc). Some higher expenses (e.g. kids) are perhaps hard to avoid – but think its important to try to increase savings over time (to ensure human capital returns are reflected in savings, not just expenditures). An excellent book on this is by Jane Bryant Quinn – Smart and Simple Financial Strategies for Busy People.

An overall investment framework helps (much): Having a clear (and relatively simple) framework can help (significantly) in guiding overall investment decisions. For me this has been the Fama-French three factor model as reflected in my investment policy. This takes out a lot of the noise (e.g. marketing hypes, continuous new product offerings etc) from the investment process – and helps to focus on what’s important. The Fama-French model is not perfect, but IMO is better than alternatives. While my factor exposure targets (and intl split) has not changed in 10 years, the underlying funds I use have – to better match my factor load targets in a more cost effective way. For example, when I started there were no readily available intl. value and intl. small cap ‘index’ funds (and switched over from Dodge and Cox Intl. and Vanguard International Explorer when these became available). An excellent article on the Fama-French model is Common Risk Factors in the Returns on Stocks and Bonds.

Fixed income for downside protection: 2008 reiterated this point (to me) very well. US treasuries provided the greatest protection when needed most. Take risk with equities – use fixed income for downside protection – at least this is my approach. IMO useful reading of this is in David Swensen’s books (Pioneering Portfolio Management, and Unconventional Success).

Do you own research: Don’t believe everything you read, do your own research and analysis. This helps given a much better feel of the robustness or paucity of what you read. And helps strengthen you own decisions and ability to stay the course (data is now more readily available via Ken French’s website, and by some of the index providers such as S&P, Russell, MSCI, CRSP).

Final word of thanks. I have learned much from the M* board and this board over the last 10 years. While not an exhaustive list in any way this includes – from Larry Swedore, Bill Bernstein, Rick Ferri, Taylor Larimore, Jeff Troutner – many thanks. I have also learned from debates with Eric (in his various forms e.g. Smallhi) – it is debate which helps test and refine ideas [I think we agree on the most important things:)]. And thanks to Alex and Larry for setting up and maintaining this site.

Best for the new year, and the next 10!

Robert
.
livesoft
Posts: 85972
Joined: Thu Mar 01, 2007 7:00 pm

Re: 10 Year Retrospective

Post by livesoft »

That's great! Thanks for posting!

I have followed a similar Fama-French-inspired small-cap and value-tilted strategy, so you know I'm liking it.

But what I really wanted to know was:

1. How do YOU PERSONALLY do your rebalancing? And did you rebalance significantly in March 2009?
2. How do YOU PERSONALLY do your tax-loss harvesting? How big a loss do you need?

Thanks!
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
Noobvestor
Posts: 5944
Joined: Mon Aug 23, 2010 1:09 am

Re: 10 Year Retrospective

Post by Noobvestor »

Thank you, Robert. It is great to see your update, and to hear your thoughts a decade in. You have contributed a lot to this forum. Anyone who hasn't should read and bookmark this thread of yours in particular, IMO: http://www.bogleheads.org/forum/viewtopic.php?t=7353
"In the absence of clarity, diversification is the only logical strategy" -= Larry Swedroe
User avatar
Topic Author
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: 10 Year Retrospective

Post by Robert T »

livesoft wrote:But what I really wanted to know was:

1. How do YOU PERSONALLY do your rebalancing? And did you rebalance significantly in March 2009?
2. How do YOU PERSONALLY do your tax-loss harvesting? How big a loss do you need?

Thanks!
Livesoft,

An attempt at your two questions.

1. How do YOU PERSONALLY do your rebalancing? And did you rebalance significantly in March 2009?

For most of the last 10 years: New monthly savings have gone to most underweighted asset classes, then I used a rebalancing band approach for any sell/buy rebalancing – along the lines of the 5x25 guide advocated by Larry. i.e. if allocation gets more than about 5% in absolute terms or 25% in relative terms (whichever comes first) then rebalance. I rebalanced significantly between stocks and bonds at end October 2008 (buying more stocks). As all stock funds had declined there was not much ‘relative’ rebalancing across stock funds needed – except for EM and Intl. SC which has declined by more. I did not rebalance significantly again in March as perhaps was not past threshold again. So against an annually rebalanced benchmark, would have lost more in 2008 – as I rebalanced into continually declining stocks. More recently I have given more consideration to a valuation based approach to rebalancing. Within the rebalancing bands I overweight slightly those asset classes with highest expected return based on GMO expected returns (at aggregate US:EAFE:EM:Bond level). Back tests suggest some added value. Time will tell.

2. How do YOU PERSONALLY do your tax-loss harvesting? How big a loss do you need?

For tax-loss harvesting – I have tended to focus on the larger fund holdings in my portfolio (and those for which there are reasonable substitutes) – and it’s a balance between keeping transaction costs low and the size (value) of the tax loss harvested (both lot size and the magnitude of decline matter). I have not used any hard rule (e.g. when this gets to x % of fund holding then harvest the losses). Simply when there are significant market declines (over extended periods), I check to see any harvesting opportunities. Perhaps I should give more thought to clearer rules of thumb.

Robert
.
User avatar
Taylor Larimore
Posts: 32839
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

Re: 10 Year Retrospective

Post by Taylor Larimore »

Robert:

Thank you for an excellent retrospective.

I am glad I am one of many who helped you along the way. Thanks for remembering.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
livesoft
Posts: 85972
Joined: Thu Mar 01, 2007 7:00 pm

Re: 10 Year Retrospective

Post by livesoft »

Robert T wrote:An attempt at your two questions.
....
Thank you for answering my questions. I appreciate your responses and contributions to the forum.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
Noobvestor
Posts: 5944
Joined: Mon Aug 23, 2010 1:09 am

Re: 10 Year Retrospective

Post by Noobvestor »

Robert T wrote:Within the rebalancing bands I overweight slightly those asset classes with highest expected return based on GMO expected returns (at aggregate US:EAFE:EM:Bond level). Back tests suggest some added value. Time will tell.
Interesting. So how much are you overweight EM/EAFE right now? :)
"In the absence of clarity, diversification is the only logical strategy" -= Larry Swedroe
STC
Posts: 415
Joined: Wed Nov 14, 2012 8:22 am

Re: 10 Year Retrospective

Post by STC »

Great post, and thanks. I have a very similar portfolio, only I add 10% REIT 1/2 US, 1/2 Intl. otherwise, identical!

I'm glad to see it worked out well for you during the "lost decade"
User avatar
Topic Author
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: 10 Year Retrospective

Post by Robert T »

.
Noobvestor wrote:Interesting. So how much are you overweight EM/EAFE right now? :)
It works out at 45:39:16 for US:EAFE:EM equity - 'overbalanced' (as Bernstein calls it in his first book) to EAFE and EM based on GMO expected return but within the overall portfolio 5x25 rebalancing bands. I review monthly with new savings to keep within bands. This approach 'hurt' relative returns in 2011, but helped in 2012. Time will tell.

Robert
.
Snowjob
Posts: 1650
Joined: Sun Jun 28, 2009 10:53 pm

Re: 10 Year Retrospective

Post by Snowjob »

Agreed with the other responders, I have enjoyed your work and the debates with Small-Hi over the last 5 years that I've been reading the forums.

Given your meticulus calculations my assumption is you stripped out from your portfolio returns, the value of new contributions and that adding these back into the mix would have bumped up absolute growth of the porfolio by quite a factor.

Also, care to give us a preview of your expected returns for the next 10 years based on Asset Class mix as of Jan 1.2013?

Best of luck!
User avatar
k66
Posts: 504
Joined: Sat Oct 27, 2012 1:36 pm

Re: 10 Year Retrospective

Post by k66 »

Thank you Robert. Your data and experience are a great beacon to some of us newer followers. Regarding some of your established limits or expectations, for example the maximum 1 year loss of 30 to 35%, how were they initially produced? I presume from historical data, but did you then make a probabilistic calculation, or simply estimate? Also, did you have any expectation or estimate for the average period (10 year) inflation rate?

Kevin
LOSER of the Boglehead Contest 2015 | lang may yer lum reek
User avatar
Jay69
Posts: 1801
Joined: Thu Feb 17, 2011 8:42 pm

Re: 10 Year Retrospective

Post by Jay69 »

Robert

Thanks for the post, very nice.

May I ask your age and did you hold at 25/75% over the entire 10 years?

Going to change your 25/75% going forward.

I prefer to hold more of a static AA myself.
"Out of clutter, find simplicity” Albert Einstein
User avatar
ofcmetz
Posts: 2465
Joined: Tue Feb 08, 2011 7:09 pm
Location: Louisiana

Re: 10 Year Retrospective

Post by ofcmetz »

Thanks for posting your IPS and investing results from the last ten years. It was a very interesting and informative post.
Never underestimate the power of the force of low cost index funds.
ScottW
Posts: 165
Joined: Tue Mar 25, 2008 8:13 am

Re: 10 Year Retrospective

Post by ScottW »

Thanks for all the posts and careful analysis you've done over the years. Although you haven't posted as much over the past year or so, I got quite a bit out of the regular (and civil) debates you used to have with SmallHi and others back when I first joined the board.
User avatar
ddb
Posts: 5511
Joined: Mon Feb 26, 2007 11:37 am
Location: American Gardens Building, West 81st St.

Re: 10 Year Retrospective

Post by ddb »

Robert T wrote:.Actual over last ten years
Annualized Return 2003-12 = 10.1%
Largest one year loss (2008) = 28.7%
Is your annualized return figure time-weighted or dollar-weighted?

- DDB
"We have to encourage a return to traditional moral values. Most importantly, we have to promote general social concern, and less materialism in young people." - PB
reggiesimpson
Posts: 1697
Joined: Mon Dec 19, 2011 11:47 am

Re: 10 Year Retrospective

Post by reggiesimpson »

Impressive. Congratulations.
nepats
Posts: 396
Joined: Mon Jul 20, 2009 11:56 am

Re: 10 Year Retrospective

Post by nepats »

Thanks for sharing. There is hope, j/k
DualIncomeNoDebt
Posts: 436
Joined: Wed Jul 18, 2012 3:38 am

Re: 10 Year Retrospective

Post by DualIncomeNoDebt »

Nice post thank you. Your first two points about having a simple plan and then following it are key. So nice to tune out the financial talking heads, frankly necessary to keep one sane.
User avatar
Topic Author
Robert T
Posts: 2803
Joined: Tue Feb 27, 2007 8:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: 10 Year Retrospective

Post by Robert T »

.
Snowjob –
  • Yes, returns strip out the value of new contributions. On future expected returns – to be consistent, I should be expecting about a 5% return over the next 10 years, which together with the 10.1% over the last decade would give a 7.5% annualized return for the 20 years (my long-term expectation). So expecting less than 7.5% returns over next 10 years (i.e. will not be suprised if it is less than 7.5%).
K66 –
  • On the expected 1 year loss estimate – my assumption is a 7.5% return, with a standard deviation of 14 (reflective of long-term historical performance). If financial returns were normally distributed this implies about a 7% loss every 6 years (1 standard deviation), a 21% loss every 44 years (2 standard deviations), and a 35% percent loss every 740 years (3 standard deviations). Clearly large losses (e.g. -35%) happen more frequently that the normal distribution implies (with ‘fatter tails’ in the distribution in financial returns). An alternative distribution, which accounts for these ‘fatter tails’ (the more frequent occurrence of outliers) was proposed by Mandelbrot and Taleb – called fractals (e.g. see Mild vs. Wild Randomness: Focusing on those risks that matter) e.g. if a 7 percent loss is expected every 6 years, then a 14 percent loss should be expected every 16 years and 28 percent loss expected every 35 years (using a tail exponent of 1.5 re: linked article)). So while a 30-35% loss may seem highly unlikely (very infrequent) from the normal distribution (re: 1 in 740 years), in reality it has occurred more frequently (closer to the ‘fractal estimate’ of about 1 in 35 years). So I should not be ‘surprised’ with a 35% loss (larger losses are not impossible). Another estimate I use (for tracking purposes) is that advocated by Zvi Bodie presented in the chart below (still needs to be updated with 2012 returns). It shows the cumulative lower and upper (dotted) lines for which there is a 16% chance of being below or above these lines respectively (in accumulated portfolio value) based on an expected return of 7.5%, standard deviation of 14, and constant annual savings amount. There are obviously no guarantees, but think its important to try to understand some of these odds. The chart my seem oversimplified, but think it’s instructive (a more detailed analysis can be provided using the Monte Carlo simulations at Fidelity and Financial Engines to derive odds of success which is also instructive).

    Image

    On the inflation rate was about 2.5% over last 10 years (see ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt).
Jay69 -
  • Age early 40s. I have not changed my direct 75:25 stock:bond allocation over the last 10 years (its remained constant). However I have had a mortgage over this period which is a negative bond position. When I started, the mortgage was larger than my 25% bond holding suggesting I was leveraged at the start (overall having a negative bond holding). The magnitude of leverage declined with each mortgage payment and monthly investment. The mortgage is currently less that my bond holding (but still positive) so implicitly have a lower than 25 percent bond holding overall. In this sense, my overall bond holding has increased from a negative position, to positive over the last 10 years, and will only be at 25 percent when I pay-down (off) the mortgage - still a few years. I don’t plan on changing my direct 75:25 stock:bond holding going forward.
DDB –
  • My annualized returns presented earlier are time-weighted, for comparison purposes (can be compared against other portfolios e.g. those in madsinger’s monthly reports, with DFA portfolio etc, and my 7.5% expected return which is also time-weighted).
Best to all,

Robert
.
User avatar
Federal Agent
Posts: 134
Joined: Fri Feb 15, 2008 8:14 pm

Re: 10 Year Retrospective

Post by Federal Agent »

Hi Robert,

I don't think I have seen your Ideal/IPS portfolio in the Boglehead format.... It would be insightful to know about your asset LOCATION, in addition to your prescribed asset allocation.

Taxable
xx% cash (for investing – do not include emergency funds)
xx% fund name (ticker symbol)
xx% fund name (ticker symbol)

401k
xx% fund name (ticker symbol)
xx% fund name (ticker symbol)
xx% fund name (ticker symbol)

Roth IRA
xx% fund name (ticker symbol)
xx% fund name (ticker symbol)

IRA
xx% fund name (ticker symbol)
xx% fund name (ticker symbol)

100% Total Portfolio

Another question: If you WERE all "taxable," can you show us how you would have built a very long-term, tax efficient, globally diversified portfolio using only Vanguard index funds/Vanguard ETFs.

Would you reduce number of asset class holdings or allocation amounts of your small/value holdings to minimize the potential tax liabilities of re-balancing? (knowing of course this would reduce your FF factor loadings); Say qualified dividend tax treatment were eliminated (or remained ongoing), would this change the way you'd invest, given no tax friendly space?

Taxable Account - Vanguard only investments
xx% cash (for investing – do not include emergency funds)
xx% Vanguard fund name or Vanguard ETF (ticker Symbol)
xx% Vanguard fund name or Vanguard ETF (ticker Symbol)
xx% Vanguard fund name or Vanguard ETF (ticker Symbol)
100% Total Portfolio

Thanks!
User avatar
serbeer
Posts: 1304
Joined: Fri Dec 28, 2007 1:09 pm

Re: 10 Year Retrospective

Post by serbeer »

Robert,
thank you very much for this very interesting summary of your personal performance over the past 10 years. I also happened to have 75/25 portfolio that I also started nearly 10 years ago (2003) except that I only got my spreadsheets in sufficient order to track XIRR-based ROR by 2009--mostly thanks to help of folks on this forum and ultimatum from my wife that she will no longer support investing into market unless I can track and show her total ROR across all our accounts :) By the way, it came to 9.4% overall return for the past 3 years in which I tracked it.

While reading your post, I could not help but wonder about a few things which I may consider tracking/adding to my own financial spreadsheet. Could you elaborate on the following items please?

1) How did you arrive to "Estimated source of return 2003-12" numbers in general and for "Value+Size tilt" in particular? I assume you did not track exact ROR for each of these in your spreadsheet, but the total came to exactly the same number as "Actual over last ten years" return. Does it mean that you worked backward from this number somehow to come up with detailed breakdown?

2) In this post, and several previous posts, you refer to Zvi Bodie estimate for 16% deviation from average expected return. I could not find any article or formula for this calculation. Could you point me to it please? I hope it will clarify of significance of 16%, which I cannot figure out (though one line hints that it is "about 60% of target"), unless this number is a function of expected return and standard deviation picked for your AA.

3) Why do you consider these 16% odds important in portfolio performance tracking?

4) For the chart itself, have you picked initial amount of investment and charted exact 7.5% return curve while building hypothetical progress based on growth of the same initial amount based on of actual year by year returns of your portfolio?

5) Myself, I do not have a similar chart, but thought it may be a good idea to build it, hence the questions. The chart I currently use shows stacked columns of annual liquid net worth snapshots with component of each column being last year portfolio value, annual portfolio growth, and total annual contributions. So it gives very good visual information of past performance and overall trend but does not show "progress toward target", and that was because annual contribution changes quite a bit over the years so I could not come up with some "expected" line to plot. Perhaps I should consider doing what you do (as per question #4). But I cannot figure out how to interpret such graph. For example, what if due to higher (or lower) than expected annual contributions in some of the years, rather than due to ROR of the portfolio, I am advancing (or faling behind) to my target much faster (or slower) than expected? Your type of chart would not indicate it, would it?

Thank you!
User avatar
k66
Posts: 504
Joined: Sat Oct 27, 2012 1:36 pm

Re: 10 Year Retrospective

Post by k66 »

Thank you Robert - the Mandelbrot article was interesting and I was not aware that he has also delved into the nature of market forces and portfolio theory. The paper actually raises more questions than it answered, but isn't that always the case!? I will track down some of the reference material listed for further interest.

---

Serbeer - I can perhaps shed some light on the 16% limits presented. The 50% growth line occurs when growth is maintained at the mean value with a 50% likelihood that growth could be higher than the line and, conversely, a 50% likelihood that it is less.

The 16% lines are actually the +/- 1sigma (1 standard deviation) lines - where in a normally distributed population 66.7% of the population exists within 1sigma of the mean. The remaining 33.3% exists either above +1sigma and below -1sigma. Assuming that the remaining distributions are equal to the top and bottom, there would be 16.7% above +1sigma and the final 16.7% below -1sigma.

One would thus expect to be bounded between the two dashed lines 66.7% of the time, and be either higher or lower than the boundaries 33.3% of the time (16.7% outside either extreme).
LOSER of the Boglehead Contest 2015 | lang may yer lum reek
User avatar
serbeer
Posts: 1304
Joined: Fri Dec 28, 2007 1:09 pm

Re: 10 Year Retrospective

Post by serbeer »

k66 wrote: ...
---
Serbeer - I can perhaps shed some light on the 16% limits presented. The 50% growth line occurs when growth is maintained at the mean value with a 50% likelihood that growth could be higher than the line and, conversely, a 50% likelihood that it is less.

The 16% lines are actually the +/- 1sigma (1 standard deviation) lines - where in a normally distributed population 66.7% of the population exists within 1sigma of the mean. The remaining 33.3% exists either above +1sigma and below -1sigma. Assuming that the remaining distributions are equal to the top and bottom, there would be 16.7% above +1sigma and the final 16.7% below -1sigma.

One would thus expect to be bounded between the two dashed lines 66.7% of the time, and be either higher or lower than the boundaries 33.3% of the time (16.7% outside either extreme).
Thank you k66. I was under impression that 68.xx% of population exists within 1 sigma given normal distribution (68-95-99.7 rule). But perhaps another distribution is modeled? I hope Robert can clarify.
User avatar
k66
Posts: 504
Joined: Sat Oct 27, 2012 1:36 pm

Re: 10 Year Retrospective

Post by k66 »

Serbeer - You are correct, it is actually 68.2% for a normally distributed population... sometimes I just like to express simple fractions as approx values (1/3, 1/6, etc).

The more interesting point is that, if as Mandelbrot et al suggest, the distribution is really not normal; then what would be the appropriate percentages associated with the tails? I would also speculate that the distribution has unequal tails (or there is some kurtosis/skew in the population and one tail is fatter than the other).

-Kevin
LOSER of the Boglehead Contest 2015 | lang may yer lum reek
haban01
Posts: 677
Joined: Thu Mar 01, 2007 8:55 pm
Location: Wisconsin

Re: 10 Year Retrospective

Post by haban01 »

Robert, I believe that we sat together and had dinner at Bogleheads reunion in 2007 in Alexandria? Is that correct?
Eric | | "Stay the Course" | "Press on Regardless"
gt4715b
Posts: 533
Joined: Mon Jun 11, 2007 10:29 am

Re: 10 Year Retrospective

Post by gt4715b »

How are you calculating the sources of returns, particularly the international diversification part? I would like to add this capability to my portfolio tracking spreadsheet.
Post Reply