Author Topic: In retirement the PP seems to work, what am I missing??  (Read 19984 times)

effigy98

  • Pencil Stache
  • ****
  • Posts: 555
In retirement the PP seems to work, what am I missing??
« on: January 05, 2016, 09:17:24 PM »
I'm not talking about the accumulation phase, only after FI.

I find so many counter opinions on this board that are arguing against the PP portfolio. A slightly modified one called the Golden Butterfly seems to be really good in retirement. Am I missing something and this is not the case?

Talks about SWR and how it impacts your money over time
http://portfoliocharts.com/2015/11/17/how-safe-withdrawal-rates-work/

Golden Butterfly
http://portfoliocharts.com/portfolio/golden-butterfly/

Crushtheturtle

  • 5 O'Clock Shadow
  • *
  • Posts: 51
Re: In retirement the PP seems to work, what am I missing??
« Reply #1 on: January 06, 2016, 05:55:14 AM »
"In retirement, the PP seems seemed to work."

Fixed

And in a rising interest rate environment, high bond allocations may not work out well.

Louisville

  • Pencil Stache
  • ****
  • Posts: 545
Re: In retirement the PP seems to work, what am I missing??
« Reply #2 on: January 06, 2016, 07:15:38 AM »
The peepee seems to work? So, no erectile dysfunction in retirement? That's good.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #3 on: January 06, 2016, 09:17:04 AM »
@effigy98: You're not missing anything -- the Permanent Portfolio is an excellent retirement portfolio. 

By holding a good amount of treasuries it will never be the best portfolio when rates rise.  But by simultaneously holding stocks, gold, and cash, it will also never be the worst.  Compare the worst times of the PP to that of the stock market alone:




Note that the PP did consistently well even when interest rates and inflation skyrocketed during the 70's, gold tanked in the 80's and 90's, and stocks stagnated in the 2000's.  That's diversification and wealth preservation at work.  Because of the way SWRs work (as outlined in your link), it's that stability along with the good (if not world-beating) returns that makes it work so well in retirement. 

FWIW, I am personally FIRE and the PP is my portfolio of choice.  I'm very happy with it.

As an aside, many people simply dislike long term treasuries or gold for a variety of reasons and they naturally will dislike the PP regardless of its performance.  I actually have no problem with that.  There are many portfolios that perform similarly well by diversifying beyond simple stocks and bonds, and perhaps one of them will be more appealing.  It's not about the Permanent Portfolio vs the stock market or about finding the single "ideal" asset allocation to rule them all.  It's more about understanding the benefits of diversification and thinking about investing in holistic terms beyond simply maximizing average returns.  Find an asset allocation you'll be comfortable with both in good times and bad and are unlikely to ever sell low, and you'll be way ahead of most investors. 
« Last Edit: January 06, 2016, 09:40:18 AM by Tyler »

zz_marcello

  • 5 O'Clock Shadow
  • *
  • Posts: 52
Re: In retirement the PP seems to work, what am I missing??
« Reply #4 on: January 06, 2016, 09:44:24 AM »
In the 70's and also 00's Gold rocketed both times.
In the 70s the move was exaggerated by the artificial $30 per ounce price fix. (so Gold could make more than 2000% from 1972 to 1980; offsetting both the stock market and bond market crashes). In my personal opinion its not likely that Gold will do that again in this magnitude in case the yield curve turns around in the next 10 years; because in the long term Gold is "only" an inflation protection and doesn't generate any yield.

Portfoliocharts in general is a beautiful and outstanding work and I'm thankful that it exists! THANK YOU!
What would further improve it is a data set, that would include the 1960's. (and of course even better the period since 1900).
The FIRE years ~1902 and ~1966 where the most challenging, because both where the start of long inflationary periods with initially very high stock valuations that came down in a dreading 15+ period to record lows.
Currently the stock market valuations (US CAPE25+) are as high as during those starting years
This is why the argumentation "a x% withdrawal rate is in general xx% safe" is not valid for me in this environment.

In my view the argumentation has to be:
"How high was the probability of a safe withdrawal rate in those periods, when US stock market valuations where at CAPE25+ and bond yields where very low"?

In my personal view I belief that investors who only rely on US stocks and treasury bonds will have a very high "chance" of seeing a significant drop in the next few years and Gold will not offset this enough.
Things like US Real Estate Rentals, US REITS, lower valuated Asian emerging market stocks and also new tools like consumer debt lending (Lending Club) and private equity investment (Fundrise) are for me personally tools to lower overall portfolio volatility.

Have a nice week!
« Last Edit: January 06, 2016, 10:34:58 AM by zz_marcello »

effigy98

  • Pencil Stache
  • ****
  • Posts: 555
Re: In retirement the PP seems to work, what am I missing??
« Reply #5 on: January 06, 2016, 09:52:18 AM »
Thank you Tyler for the information, it is very helpful to get a strong opinion on it one way or another.

The only thing I don't like about the permanent portfolio is holding gold is sketchy as GLD does not seem to be required to back their gold and I do not trust holding physical "offsite", so I would feel I would need possession of the physical gold, so re-balance can be a pain and there is weird tax implications because its considered a collectible. Another problem with gold is places like APMEX charge a premium over the spot price so it's not as good of deal as it looks on the surface. This one thing drives me towards another portfolio instead, but at the same time I do not want to leave money on the table.

The other question is does portfolio charts cover dividend reinvests. One of my favorite ETFs (VYM (or its holdings since it is newish) looks just average on a lot of back testers, but when I actually find a good dividend back tester, it looks amazing for total returns. I really like the thought of dividends during the withdraw phase as it "feels" like I can avoid touching the principle for the most part and let the price appreciation of the stock grow.

What are your thoughts on these points?

Louisville and Crushtheturtle, with my OUTRAGEOUS optimism. I will have a functioning PP in retirement... However, just to be safe, you should budget some pills into your SWR.

marcello, I really like the vanguard REIT, I also have a rental property scooped up during the crash. I have avoided Lending Club because I heard the taxes are a bit of a pain, have they fixed that with a better statement yet? Tax time is already a stressful enough situation for us. I really like the idea for both Fundrise and Lending Club.
« Last Edit: January 06, 2016, 10:06:11 AM by effigy98 »

onlykelsey

  • Handlebar Stache
  • *****
  • Posts: 2167
Re: In retirement the PP seems to work, what am I missing??
« Reply #6 on: January 06, 2016, 09:53:51 AM »

Louisville and Crushtheturtle, with my OUTRAGEOUS optimism. I will have a functioning PP in retirement... However, just to be safe, you should budget some pills into your SWR.

+1.

No thought on the portfolio allocation, although it made me realize I have an instinctual gut aversion to holding gold/commodities when I looked up what it was and immediately wanted to dismiss it.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #7 on: January 06, 2016, 10:00:31 AM »
The only thing I don't like about the permanant portfolio is holding gold is sketchy as GLD does not seem to be required to back their gold and I do not trust holding physical "offsite", so I would feel I would need possession of the physical gold, so re-balance can be a pain and their are weird tax implications because its considered a collectable. Another problem with gold is places like APMEX charge a premium over the spot price so it's not as good of deal as it looks on the surface. This one thing drives me towards another portfolio instead, but at the same time I do not want to leave money on the table.

The other question is does portfolio charts cover dividend reinvests. One of my favorite ETFs (VYM (or its holdings since it is newish) looks just average on a lot of backtesters, but when I actually find a good dividend backtester, it looks amazing for total returns.

What are your thoughts on these points?

Regarding gold (and the PP in general), I recommend this book. There's an entire chapter on ways to buy and hold gold along with all of the various tradeoffs.

And yes, all Portfolio Charts data includes reinvested dividends. 
« Last Edit: January 06, 2016, 10:03:21 AM by Tyler »

effigy98

  • Pencil Stache
  • ****
  • Posts: 555
Re: In retirement the PP seems to work, what am I missing??
« Reply #8 on: January 06, 2016, 10:08:28 AM »
Include dividend reinvest, nice!!

Funny you should mention that book. I am about 1/4 of the way thru just kindle downloaded it last week.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #9 on: January 06, 2016, 10:13:11 AM »
@zz_marcello: No worries, and thanks!  I touch on your data timeframe concern a bit here, but I applaud your thoughtfulness on your own portfolio and am confident you're doing fine.  I totally agree that there's a lot more to retirement investing than blindly following a SWR calculated over any timeframe. 

Scandium

  • Magnum Stache
  • ******
  • Posts: 2827
  • Location: EastCoast
Re: In retirement the PP seems to work, what am I missing??
« Reply #10 on: January 06, 2016, 10:13:23 AM »
Gold and bonds have done really well last several decades (times of high inflation and dropping interest rates), so high allocations to those assets would do well when only going back to 1972. How did they do before then? Also, we are now (probably) going the other direction; rising rates and low inflation. At least for a while. What will this do to the PP performance?

I'm all for more diversification, but gold never seemed very appareling to me. I'd rather do more international stocks/bonds, REITS, and properties.

zz_marcello

  • 5 O'Clock Shadow
  • *
  • Posts: 52
Re: In retirement the PP seems to work, what am I missing??
« Reply #11 on: January 06, 2016, 10:37:46 AM »
@zz_marcello: No worries, and thanks!  I touch on your data timeframe concern a bit here, but I applaud your thoughtfulness on your own portfolio and am confident you're doing fine.  I totally agree that there's a lot more to retirement investing than blindly following a SWR calculated over any timeframe.

Hi Tyler,

I read all your work, also your point with the data set.
Thanks again for all that you put together!
It helped me a lot!
« Last Edit: January 06, 2016, 10:39:36 AM by zz_marcello »

Crushtheturtle

  • 5 O'Clock Shadow
  • *
  • Posts: 51
Re: In retirement the PP seems to work, what am I missing??
« Reply #12 on: January 06, 2016, 01:14:20 PM »

Louisville and Crushtheturtle, with my OUTRAGEOUS optimism. I will have a functioning PP in retirement... However, just to be safe, you should budget some pills into your SWR.

+1.

No thought on the portfolio allocation, although it made me realize I have an instinctual gut aversion to holding gold/commodities when I looked up what it was and immediately wanted to dismiss it.

What do you mean I should budget pills into my SWR? You mean medical pills?

If that's the case, I just thought of a place you can store your physical gold.


MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 6635
Re: In retirement the PP seems to work, what am I missing??
« Reply #13 on: January 07, 2016, 02:23:51 PM »
Gold averaged 64% per year between 1972-1974.  $100 in gold became $438 in told in that short time period, which should tell you something is up.  In Aug 1971, the U.S. dropped the gold standard, letting the price of gold float (see https://en.wikipedia.org/wiki/Nixon_Shock).  Notice how the charts above all include this time frame - it's very critical to the case for gold.

Another perspective: 1972-2015, gold averaged +7.1% a year.  But cut away those early years, and you are left with 1975-2015 where gold averaged +3.8% a year.  Half of gold's success comes from those 3 years.  Maybe you still want less volatility in your portfolio and you still include gold, but go in with eyes open on those critical 1972-1974 years.

Retire-Canada

  • Walrus Stache
  • *******
  • Posts: 8690
Re: In retirement the PP seems to work, what am I missing??
« Reply #14 on: January 07, 2016, 04:06:20 PM »
Quote
One bright spot amid the worldwide rout was the precious metals sector, including gold, which is seen as a safe haven in times of economic uncertainty. The February bullion contract on commodity markets rose by $16.50, to $1,108.40 US an ounce.

Read this ^^ just now in reference to the markets tumbling around the world today.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #15 on: January 07, 2016, 04:39:39 PM »
Gold averaged 64% per year between 1972-1974.  $100 in gold became $438 in told in that short time period, which should tell you something is up.  In Aug 1971, the U.S. dropped the gold standard, letting the price of gold float (see https://en.wikipedia.org/wiki/Nixon_Shock).  Notice how the charts above all include this time frame - it's very critical to the case for gold.

Another perspective: 1972-2015, gold averaged +7.1% a year.  But cut away those early years, and you are left with 1975-2015 where gold averaged +3.8% a year.  Half of gold's success comes from those 3 years.  Maybe you still want less volatility in your portfolio and you still include gold, but go in with eyes open on those critical 1972-1974 years.

MustacheAndaHalf nails it.  The "case" for gold relies on 3 years in the history of investing – in a data set that's already about 1/10 the size of other asset classes.  It's intriguing, but it smells too much like a one-time adustment (getting to proper value after letting it float) for me to think the data is at all useful for retirement planning.

Tyler, can you generate those same charts with a custom start date to address these questions?  I took a quick look at your site and I didn't see an obvious way to change what data was used.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: In retirement the PP seems to work, what am I missing??
« Reply #16 on: January 07, 2016, 04:55:09 PM »
Gold averaged 64% per year between 1972-1974.  $100 in gold became $438 in told in that short time period, which should tell you something is up.  In Aug 1971, the U.S. dropped the gold standard, letting the price of gold float (see https://en.wikipedia.org/wiki/Nixon_Shock).  Notice how the charts above all include this time frame - it's very critical to the case for gold.

Another perspective: 1972-2015, gold averaged +7.1% a year.  But cut away those early years, and you are left with 1975-2015 where gold averaged +3.8% a year.  Half of gold's success comes from those 3 years.  Maybe you still want less volatility in your portfolio and you still include gold, but go in with eyes open on those critical 1972-1974 years.

MustacheAndaHalf nails it.  The "case" for gold relies on 3 years in the history of investing – in a data set that's already about 1/10 the size of other asset classes.  It's intriguing, but it smells too much like a one-time adustment (getting to proper value after letting it float) for me to think the data is at all useful for retirement planning.

Tyler, can you generate those same charts with a custom start date to address these questions?  I took a quick look at your site and I didn't see an obvious way to change what data was used.

Why not just look at the chart above and only look at the post 1972 years?

This chart: https://portfoliocharts.files.wordpress.com/2015/08/permanent-portfolio-cagr-pixel-21.jpg

And ignore the first two rows.

You can make your own allocation, but it'll show all years 72-present: http://portfoliocharts.com/portfolio/pixel/

But, again, just ignore those.  It doesn't affect all the other rows (but obviously does the overall numbers cited in the summary box)
« Last Edit: January 07, 2016, 04:57:53 PM by arebelspy »
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: In retirement the PP seems to work, what am I missing??
« Reply #17 on: January 07, 2016, 05:00:05 PM »
On topic, I'm not a fan of the PP for early retirees.  The PP seems too risky to me, long term (i.e. a 60 year ER). I don't think the real return will be enough long term, because I think underperforming parts of the PP will drag down your return, so I see it as riskier long term.  For short term wealth stability, it seems pretty solid.  For late (normal aged) retirees, it seems like a fine option.  YMMV.

I very much agree with Tyler's summary paragraph from earlier:
Quote
As an aside, many people simply dislike long term treasuries or gold for a variety of reasons and they naturally will dislike the PP regardless of its performance.  I actually have no problem with that.  There are many portfolios that perform similarly well by diversifying beyond simple stocks and bonds, and perhaps one of them will be more appealing.  It's not about the Permanent Portfolio vs the stock market or about finding the single "ideal" asset allocation to rule them all.  It's more about understanding the benefits of diversification and thinking about investing in holistic terms beyond simply maximizing average returns.  Find an asset allocation you'll be comfortable with both in good times and bad and are unlikely to ever sell low, and you'll be way ahead of most investors.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #18 on: January 07, 2016, 05:04:34 PM »
Gold averaged 64% per year between 1972-1974.  $100 in gold became $438 in told in that short time period, which should tell you something is up.  In Aug 1971, the U.S. dropped the gold standard, letting the price of gold float (see https://en.wikipedia.org/wiki/Nixon_Shock).  Notice how the charts above all include this time frame - it's very critical to the case for gold.

Another perspective: 1972-2015, gold averaged +7.1% a year.  But cut away those early years, and you are left with 1975-2015 where gold averaged +3.8% a year.  Half of gold's success comes from those 3 years.  Maybe you still want less volatility in your portfolio and you still include gold, but go in with eyes open on those critical 1972-1974 years.

MustacheAndaHalf nails it.  The "case" for gold relies on 3 years in the history of investing – in a data set that's already about 1/10 the size of other asset classes.  It's intriguing, but it smells too much like a one-time adustment (getting to proper value after letting it float) for me to think the data is at all useful for retirement planning.

Tyler, can you generate those same charts with a custom start date to address these questions?  I took a quick look at your site and I didn't see an obvious way to change what data was used.

Why not just look at the chart above and only look at the post 1972 years?

This chart: https://portfoliocharts.files.wordpress.com/2015/08/permanent-portfolio-cagr-pixel-21.jpg

And ignore the first two rows.

You can make your own allocation, but it'll show all years 72-present: http://portfoliocharts.com/portfolio/pixel/

But, again, just ignore those.  It doesn't affect all the other rows (but obviously does the overall numbers cited in the summary box)

There are lots of places to get the annual returns of gold.  The chart you link shows that the PP returns between 3 and 6% for all 13+ year periods regardless of starting date.  I guess that's useful for something, but not really to me or this topic, and certainly not useful for the question of withdrawal rates.

I'm specifically interested in Tyler's other tools, particularly the withdrawal rate and funnel tools.

effigy98

  • Pencil Stache
  • ****
  • Posts: 555
Re: In retirement the PP seems to work, what am I missing??
« Reply #19 on: January 07, 2016, 05:07:25 PM »
This was a good site I found for answering most of my gold re-balance and holding questions. I think if I switch over to the golden butterfly, will hold half physical, half ETF (to make it easier to re balance).

http://www.crawlingroad.com/blog/2009/10/13/permanent-portfolio-25-gold-allocation-faq/

You guys make good points, I am interested in the charts after you remove those wacky years and start with something like 1976 on. I am mostly interested in the SWR phase more so than accumulation phase. I think the roller coaster volatility in non PP portfolios is probably better during accumulation, but no way I would stomach that in ER when I do not think I will go back to making the insane tech pay I can now as my skills will atrophy and it took me years and years to develop them and keep current. I never want to go back to work in tech once I ER, it is office space soul crushing.
« Last Edit: January 07, 2016, 05:17:06 PM by effigy98 »

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #20 on: January 07, 2016, 07:51:00 PM »
I must say, this is the first time I've been critiqued for basing an analysis on too much data.  ;)  But seriously, I get it.  The issue of gold behaving exuberantly in the early years and potentially skewing the data concerned me as well until I also looked deeper. 

When the gold standard was eliminated, gold did go on a huge run for three years before correcting in 1975 (falling 28%).  So for argument's sake (and to be conservative starting in a big negative year for gold) let's look at the data starting in 1975.



Arebelspy is correct -- the way the chart works, changing the start date leaves all of the individual cells completely unaffected.  In fact, the primary purpose of the "Pixel" chart is to see beyond a potentially deceptive average based on a single arbitrary time frame.  It lets you study the big picture of every investment period at a glance.

In any case, I took the liberty of recalculating the long-term stats on the right to only look at the 1975 start date*, and the numbers are basically identical to before and within normal statistical noise.  Averages for any portfolio naturally move up and down a bit based on start year, but I wager the difference here is a lot less than many people expect with the early gold years left out.  Exclude the early 70's from stocks, for example (as far too many people do), and you conveniently miss some really horrible years and the averages look a lot more rosy. That consistency regardless of start date is one of the factors that attracted me to the PP in the first place. 

(*)BTW, if anyone wants to find the specific long-term CAGR of their own asset allocation with any start year and any duration, here's how to do it.  Fire up the Pixel calculator.  Enter your asset allocation.  Now hover your cursor over the cells of the chart.  Voila!

Regarding Safe Withdrawal Rates, I think it's important to reiterate how they work.  I use the more conservative Bengen SAFEMAX methodology which finds the single worst safe withdrawal rate over all possible starting years (that I have data for).  Removing data points can never lower the SWR.  FWIW, I double-checked the data and excluding 1972-1974 makes no difference to the SWR.  The single worst year to retire with the Permanent Portfolio was in 1980, well after the initial gold rush.

So even if you don't believe that the early 70's gold performance is likely to repeat (I certainly don't), I personally don't think starting in 1975 changes the PP outlook. 

All that said, everyone is free to make their own decision given the relevant data, which is why I launched the site in the first place.  If you simply hate gold as an asset and don't trust it, great!  You won't find any argument from me.  Browse the myriad of other well-diversified portfolios with no gold and look at their impressive performance as well.  The Permanent Portfolio is just one of many good options based on Modern Portfolio Theory.  As I tend to repeat, there's so much more to portfolio construction than calculating your percentage of stocks and bonds!

Hope that helps. 
« Last Edit: January 07, 2016, 10:28:39 PM by Tyler »

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #21 on: January 07, 2016, 09:07:48 PM »
I must say, this is the first time I've been critiqued for basing an analysis on too much data.  ;)  But seriously, I get it.  The issue of gold behaving exuberantly in the early years and potentially skewing the data concerned me as well until I also looked deeper. 

Ha!  Well, to be fair, this wouldn't be an issue if you had a longer price history for gold.  Which, to be fair, exists – the value of gold was set at $20.67 per ounce in 1834 and $35 per ounce in 1934.  The CAGR from 1972 to today is about 8.0%.  From 1934 it's about 4.3%, and from 1834 it's about 2.2%.

The issue is not the first 2-3 years of your data set.  The issue is that your data set starts at a non-market value – if gold had been freely priced from 1834, with the same CAGR, the "proper" price in 1972 would have been about $416 per ounce.  It's the difference between $416 and $35 that throws the PP off.  And gold didn't reach that long-term trend line until about 1980.


Quote
FWIW, I double-checked the data and excluding 1972-1974 makes no difference to the SWR.  The single worst year to retire with the Permanent Portfolio was in 1980, well after the initial gold rush.

That's useful information!
Of course, it probably shouldn't be surprising considering that 1980 is when the price of gold peaked for the next 25 or so years.  The issue isn't really the first 2-3 years, it's the resetting of gold to market value that actually took much longer.

Quote
So even if you don't believe that the early 70's gold rush is likely to repeat (I certainly don't), I personally don't think starting in 1975 changes the PP outlook. 

So my previous comments is why I wanted to know if we could customize the start date.  I was wrong earlier when I talked about just 1972-4 – the historical data clearly shows that gold was still catching up to inflation.

Quote
All that said, everyone is free to make their own decision given the relevant data, which is why I launched the site in the first place.  If you simply hate gold as an asset and don't trust it, great!  You won't find any argument from me.  Browse the myriad of other well-diversified portfolios with no gold and look at their impressive performance as well.  The Permanent Portfolio is just one of many good options based on Modern Portfolio Theory.  As I tend to repeat, there's so much more to portfolio construction than calculating your percentage of stocks and bonds!

Hope that helps.

I do very much appreciate your response, and the tools that you've created.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: In retirement the PP seems to work, what am I missing??
« Reply #22 on: January 08, 2016, 01:01:46 AM »
Now that gold is "caught up" to inflation, beltim, shouldn't it be able to provide the purpose of it's being in the PP in the first place?
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #23 on: January 08, 2016, 07:30:26 AM »
Now that gold is "caught up" to inflation, beltim, shouldn't it be able to provide the purpose of it's being in the PP in the first place?

In terms of providing inflation protection over the long term, probably.  But I don't think it will provide any actual real return.  Since 1980, according to two cobbled-together sets of data I pulled up, the CAGR of stocks is 10.7%, while the PP has returned 7.2% over the same period.  And when you consider that this time period has been the best for long-term treasury bonds in history, I would wager a significant amount of money that the long-term return of the PP will be significantly less than 3.5 percentage points less than stocks.

The PP certainly has less volatility than an all-stock portfolio, but I don't think its return will ever be competitive with stocks over the long term again.  I am interested in seeing how it affects SWR, but I think I'd have to go to Monte Carlo simulations in order to properly model it

brooklynguy

  • Handlebar Stache
  • *****
  • Posts: 2204
  • Age: 43
Re: In retirement the PP seems to work, what am I missing??
« Reply #24 on: January 08, 2016, 09:02:01 AM »
I think the roller coaster volatility in non PP portfolios is probably better during accumulation, but no way I would stomach that in ER

For the typical early retiree, the accumulation phase is dwarfed by the drawdown/retirement phase.  Short-term volatility is, as you say, hard for many people to "stomach" precisely because we react to it on a visceral level -- it can be gut-wrenching to watch a substantial portion of the sum total of your life savings evaporate in the relative blink of an eye.  But longevity risk, while not as emotionally scary as short-term volatility, is probably the more important risk for early retirees.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: In retirement the PP seems to work, what am I missing??
« Reply #25 on: January 08, 2016, 09:23:47 AM »
Now that gold is "caught up" to inflation, beltim, shouldn't it be able to provide the purpose of it's being in the PP in the first place?

In terms of providing inflation protection over the long term, probably.  But I don't think it will provide any actual real return.  Since 1980, according to two cobbled-together sets of data I pulled up, the CAGR of stocks is 10.7%, while the PP has returned 7.2% over the same period.  And when you consider that this time period has been the best for long-term treasury bonds in history, I would wager a significant amount of money that the long-term return of the PP will be significantly less than 3.5 percentage points less than stocks.

The PP certainly has less volatility than an all-stock portfolio, but I don't think its return will ever be competitive with stocks over the long term again.  I am interested in seeing how it affects SWR, but I think I'd have to go to Monte Carlo simulations in order to properly model it

Thanks for the thoughts.  It completely confirms the bias I already had, so I crown you genius and scholar of the day.

But longevity risk, while not as emotionally scary as short-term volatility, is probably the more important risk for early retirees.

Big +1.

A slow decline in purchasing power is much more dangerous than a (typically temporary) market crash.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #26 on: January 08, 2016, 09:55:42 AM »
But longevity risk, while not as emotionally scary as short-term volatility, is probably the more important risk for early retirees.

I totally agree.  I just do not believe that retirement longevity and low-ish volatility are mutually exclusive.  You can either address that with asset allocation (like the Golden Butterfly in the OP that matches the long-term CAGR of the stock market with less than half of the volatility) or by saving a little more and spending less than the safe WR (even with a pessimistic investing outlook, a sub-3% WR with the PP would look pretty solid IMHO).  And of course there are the myriad of other ways to fund early retirement like rentals, fun part time work, etc. 

That's why I find the topic of retirement investing so interesting.   There are many more ways to approach it than a casual read of the Trinity study might imply. 
« Last Edit: January 08, 2016, 10:10:22 AM by Tyler »

brooklynguy

  • Handlebar Stache
  • *****
  • Posts: 2204
  • Age: 43
Re: In retirement the PP seems to work, what am I missing??
« Reply #27 on: January 08, 2016, 10:14:12 AM »
I totally agree.  I just do not believe that retirement longevity and low-ish volatility are mutually exclusive.  You can either address that with asset allocation (like the Golden Butterfly in the OP that matches the CAGR of the stock market with less than half of the volatility) or by saving a little more and spending less than the safe WR (even with a pessimistic investing outlook, a sub-3% WR with the PP would look pretty solid IMHO).  And of course there are the myriad of other ways to fund early retirement like rentals, fun part time work, etc. 

That's why I find the topic of retirement investing so interesting.   There are many more ways to approach it than a casual read of the Trinity study might imply.

I totally agree with all of this too.  In addition, though, per Genius Scholar Beltim's commentary, I would caution against overlooking the inherent, categorical limitations of historical backtesting of all stripes.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #28 on: January 08, 2016, 10:21:37 AM »
I totally agree with all of this too.  In addition, though, per Genius Scholar Beltim's commentary, I would caution against overlooking the inherent, categorical limitations of historical backtesting of all stripes.

Very true.  There's definitely much more to constructing a good portfolio than looking good in backtesting. 

Retire-Canada

  • Walrus Stache
  • *******
  • Posts: 8690
Re: In retirement the PP seems to work, what am I missing??
« Reply #29 on: January 08, 2016, 10:25:54 AM »
Very true.  There's definitely much more to constructing a good portfolio than looking good in backtesting.

Aside from backtesting and monte carlo simulations how would you validate a portfolio choice without making assumptions about the future 10-40+ yrs down the road that are hard to support?

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #30 on: January 08, 2016, 10:36:48 AM »
Aside from backtesting and monte carlo simulations how would you validate a portfolio choice without making assumptions about the future 10-40+ yrs down the road that are hard to support?

That's a good question that deserves more thought than a quick post could cover. 

In the meantime, one approach I like is the one taken by both the Permanent Portfolio and the All Seasons Portfolio.  Read the work by Browne and Dalio.  The assets selected are not arbitrarily chosen to make the backtest look good and are also not based on a prediction for future growth.  They're specifically selected to address four possible economic conditions (inflation, deflation, growth, recession) that are somewhat timeless.

brooklynguy

  • Handlebar Stache
  • *****
  • Posts: 2204
  • Age: 43
Re: In retirement the PP seems to work, what am I missing??
« Reply #31 on: January 08, 2016, 10:39:33 AM »
Aside from backtesting and monte carlo simulations how would you validate a portfolio choice without making assumptions about the future 10-40+ yrs down the road that are hard to support?

Backtesting is only one tool in the arsenal for good portfolio construction and shouldn't be used in isolation.  Monte Carlo simulations are another (with their own set of flaws), and logic and reason are two more, both of which should always be employed.

You wouldn't adopt an investment strategy based on coin flips or diving messages from chicken entrails, even if it happened to backtest phenomenally.  That's an extreme example, but beltim's analysis above for why gold could be expected to perform differently in the future than in the past is a more practical example (and I don't mean to suggest that holding gold is necessarily a bad idea, but only that backtesting should be used in conjunction with, and not in lieu of, logic and reason when constructing your portfolio or adopting your investment strategy).

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: In retirement the PP seems to work, what am I missing??
« Reply #32 on: January 08, 2016, 10:41:29 AM »
Are you suggesting there should be some sort of rational explanation for why something works, beyond the fact that it has worked in the past?

Ridiculous.



(For those unacquainted with the DM thread: this is a tongue-in-cheek inside joke.  Also I don't mean to imply the PP doesn't have this--if anything it has a fairly decent explanation, IMO.)
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

Retire-Canada

  • Walrus Stache
  • *******
  • Posts: 8690
Re: In retirement the PP seems to work, what am I missing??
« Reply #33 on: January 08, 2016, 11:24:49 AM »
Backtesting is only one tool in the arsenal for good portfolio construction and shouldn't be used in isolation.  Monte Carlo simulations are another (with their own set of flaws), and logic and reason are two more, both of which should always be employed.

Presumably you used logic and reason to set the parameters of your MC simulations.

What I am getting at is there is a fairly well agreed upon idea that we shouldn't time the markets because we don't know what's going to happen even a few weeks from now let alone 20yrs+ out which is the sort of timeframe a portfolio is being built for. So can you clarify what you mean by "logic and reason" that are you suggesting gets used to evaluate which portfolio to go with?

Let's say you have 3 or 4 portfolio options you are looking at and you've confirmed they perform to your satisfaction via backtesting and MC simulation. How are you selecting between them using your logic and reason?

I ask because in practice what I have seen people do is essentially decide they can predict the future and base their decision on that or just go by gut feel which is the same, but with less deliberation.


beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #34 on: January 08, 2016, 11:35:55 AM »
I totally agree.  I just do not believe that retirement longevity and low-ish volatility are mutually exclusive.

Why?  It's both logical and empirical that higher investment vehicles have more volatility than lower investments.

Quote
You can either address that with asset allocation…by saving a little more and spending less than the safe WR (even with a pessimistic investing outlook, a sub-3% WR with the PP would look pretty solid IMHO).

Wouldn't a sub-3% WR for a 100% bond portfolio look pretty solid as well?

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #35 on: January 08, 2016, 11:46:29 AM »
In addition, though, per Genius Scholar Beltim's commentary, I would caution against overlooking the inherent, categorical limitations of historical backtesting of all stripes.

Aww, shucks..

Are you suggesting there should be some sort of rational explanation for why something works, beyond the fact that it has worked in the past?

Ridiculous.

The most hilarious illustration of this I've ever seen:
Quote
Originally, the long term model for HPQ share price was defined by the index of food without beverages ((NASDAQ:FB)) and that of rent of primary residency ((RPR)). The former CPI component led the share price by 4 months and the latter one led by 5 months. Figure 1 depicts the overall evolution of both involved indices through February 2012. Below we present five best-fit models for HPQ(t) obtained at different times:

HPQ(t) = -3.20FB(t-4) + 2.91RPR(t-5) + 3.64(t-1990) - 50.82, July 2010

HPQ(t) = -3.34FB(t-4) + 3.41RPR(t-5) + 0.51(t-1990) - 85.44, June 2011

HPQ(t) = -3.46FB(t-4) + 3.68RPR(t-5) - 0.72(t-1990) - 99.88, September 2011

HPQ(t) = -3.40FB(t-5) + 3.60RPR(t-6) - 0.57(t-1990) - 97.72, December 2011

HPQ(t) = -3.27FB(t-4) + 3.46RPR(t-5) - 0.39(t-1990) - 95.71, February 2011

where HPQ(t) is the price in U.S. dollars, t is calendar time. All coefficients have been slightly drifting but very close. This process expresses the trade-off between the linear trend in the difference between the defining CPIs and the time trend term in the above equations
from http://seekingalpha.com/article/433021-hewlett-packard-a-slight-negative-correction

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #36 on: January 08, 2016, 12:04:31 PM »
It's both logical and empirical that higher investment vehicles have more volatility than lower investments.

You can't look at individual assets in isolation.  Ditching old-fashioned "stock" and "bond" assumptions and looking at portfolios with three or more assets tends to break that conclusion.  The relationship between returns and volatility in diverse multi-asset portfolios is anything but linear.



The post on the Golden Butterfly also covers this topic.

Quote
Wouldn't a sub-3% WR for a 100% bond portfolio look pretty solid as well?

Depends on the bonds, but possibly.  With enough savings, a very wealthy cash-fiend could also never run out of money.  There's more than one way to approach retirement, and not all of them require a level of investing volatility beyond your own personal comfort zone.  We all have to find the right balance for ourselves. 
« Last Edit: January 08, 2016, 03:01:30 PM by Tyler »

brooklynguy

  • Handlebar Stache
  • *****
  • Posts: 2204
  • Age: 43
Re: In retirement the PP seems to work, what am I missing??
« Reply #37 on: January 08, 2016, 12:07:37 PM »
Aww, shucks..

Rebs crowned you with the title (but I agree that you deserve it!).

So can you clarify what you mean by "logic and reason" that are you suggesting gets used to evaluate which portfolio to go with?

I just meant that an investment strategy should never be borne solely out of backtesting, as some are (especially among the various species of market timing), without any good rational explanation to believe that their successful performance will persist into the future.  And, with respect to portfolio construction in particular, at least some of the recommended portfolios out there seem to be the result of curve-fitting from data-mining.  If you can't come up with any good explanation to support an investment strategy (beyond the fact that it worked in the past), then, in my view, it's not a good investment strategy.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #38 on: January 08, 2016, 12:09:35 PM »
It's both logical and empirical that higher investment vehicles have more volatility than lower investments.

You can't look at individual assets in isolation.  Ditching old-fashioned "stock" and "bond" assumptions and looking at portfolios with three or more assets tends to break that conclusion.  The relationship between returns and volatility in diverse multi-asset portfolios is anything but linear.



I do like that graph, but what's the justification for why these portfolios break the relationship between returns and volatility?  Now that people know there were non-linear relationships in the past, won't the market automatically correct future returns to return the relationship to linear?

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #39 on: January 08, 2016, 12:31:04 PM »
The larger question is why so many people believe the relationship is strictly linear to begin with.  IMHO, the idea that volatility is linearly proportional to returns is most often the byproduct of over-simplification.

It's true that with individual assets, the higher return options generally have more volatility.   As a result of only playing with one variable, even most 2-asset analyses will often look linear.  Since most people generally stop at "the stock market" and "the bond market", the linear conclusion is mostly baked in.  Once you start accounting for three or more volatile assets and the associated rebalancing, however, the math gets a lot more complicated.

FWIW, there's an entire field of economics dedicated to this idea.  Reducing volatility for the same return is the essence of modern portfolio theory

That's a good topic for a longer post.  Thanks for the idea. 
« Last Edit: January 08, 2016, 02:05:28 PM by Tyler »

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: In retirement the PP seems to work, what am I missing??
« Reply #40 on: January 08, 2016, 12:45:43 PM »
Essentially more risk = more return for a single asset tends to hold fairly true.  But for a portfolio with multiple semi-to-un-correlated assets, it may not necessarily be the case.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #41 on: January 08, 2016, 01:31:59 PM »
The larger question is why so many people believe the relationship is strictly linear to begin with.  IMHO, the idea that volatility is linearly proportional to returns is most often the byproduct of over-simplification.

It's true that with individual assets, the higher return options generally have more volatility.   As a result of only playing with one variable, even most 2-asset analyses will generally look linear.  Since most people generally stop at "the stock market" and "the bond market", the linear conclusion is mostly baked in.  Once you start accounting for three or more volatile assets and the associated rebalancing, however, the math gets a lot more complicated. Reducing volatility for the same return is the essence of modern portfolio theory. 

That's a good topic for a longer post.  Thanks for the idea.

You're right, I don't actually mean linear.  And I was going to say monotonic, but that's not right either based on the graphs that follow.  MPT shows that for many 2-asset combinations, there are inflection points on the risk/reward spectrum (where "risk" = volatility, a nigh-useless designation for individual investors, but useful when discussing portfolio construction).  Examples include stocks and bonds:


and US/International stocks:



and multicomponent, uncorrelated assets:


But you're suggesting that adding more asset classes can make these graphs break the relationship between risk and return.  How?  Why would that not be arbitraged away, just like an efficient market supposedly arbitrages away any publicly available information on individual stocks?

It's possible I'm not thinking about this correctly, so I really am curious - I'm not trying to say that you're wrong, I just want to know more.

Aphalite

  • Bristles
  • ***
  • Posts: 425
Re: In retirement the PP seems to work, what am I missing??
« Reply #42 on: January 08, 2016, 01:43:05 PM »
PLEASE stop associating volatility with risk, it's non-sense

Risk is not being able to maintain your lifestyle, your portfolio not keeping pace with inflation, and the aggregate of scenarios where the market does something you are not accounting for/not expecting in your inputs (ie. 4% rule which has worked for 200 years all of a sudden fails for some reason, stock market closes down for five years and there are no withdrawls possible, socialists take over and vote your wealth to be redistributed to others, Congress changes tax rules so that there is no 0% capital gains bracket which decreases the early retiree's safe withdrawal rate, etc.)

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: In retirement the PP seems to work, what am I missing??
« Reply #43 on: January 08, 2016, 02:01:06 PM »

But you're suggesting that adding more asset classes can make these graphs break the relationship between risk and return.  How?  Why would that not be arbitraged away, just like an efficient market supposedly arbitrages away any publicly available information on individual stocks?

It's possible I'm not thinking about this correctly, so I really am curious - I'm not trying to say that you're wrong, I just want to know more.

The short answer is that each portfolio has a theoretical efficient frontier, but the curve for a different set of assets will be very different. 

I'll have to think about your larger question on efficient markets for a while to formulate a clear explanation.  It's certainly a complicated topic.
« Last Edit: January 08, 2016, 11:16:43 PM by Tyler »

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Re: In retirement the PP seems to work, what am I missing??
« Reply #44 on: January 08, 2016, 02:27:02 PM »
Now that gold is "caught up" to inflation, beltim, shouldn't it be able to provide the purpose of it's being in the PP in the first place?

In terms of providing inflation protection over the long term, probably.  But I don't think it will provide any actual real return.  Since 1980, according to two cobbled-together sets of data I pulled up, the CAGR of stocks is 10.7%, while the PP has returned 7.2% over the same period.  And when you consider that this time period has been the best for long-term treasury bonds in history, I would wager a significant amount of money that the long-term return of the PP will be significantly less than 3.5 percentage points less than stocks.

The PP certainly has less volatility than an all-stock portfolio, but I don't think its return will ever be competitive with stocks over the long term again.  I am interested in seeing how it affects SWR, but I think I'd have to go to Monte Carlo simulations in order to properly model it

I think you have to be really careful with the idea that gold is a hedge against inflation. I know that is the common wisdom but I don't think it's true.

https://www.bernstein.com/Content/Research/Publications/WhatsSafe.htm

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Re: In retirement the PP seems to work, what am I missing??
« Reply #45 on: January 08, 2016, 02:29:12 PM »
I think the roller coaster volatility in non PP portfolios is probably better during accumulation, but no way I would stomach that in ER

For the typical early retiree, the accumulation phase is dwarfed by the drawdown/retirement phase.  Short-term volatility is, as you say, hard for many people to "stomach" precisely because we react to it on a visceral level -- it can be gut-wrenching to watch a substantial portion of the sum total of your life savings evaporate in the relative blink of an eye.  But longevity risk, while not as emotionally scary as short-term volatility, is probably the more important risk for early retirees.

There was a link on here where Pfau was increasing stock allocations over the course of a retirement. I think ideally you would increase your stock allocation when the markets are down. The problem with this is that its getting close to market timing.

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Re: In retirement the PP seems to work, what am I missing??
« Reply #46 on: January 08, 2016, 02:34:42 PM »
I ask because in practice what I have seen people do is essentially decide they can predict the future and base their decision on that or just go by gut feel which is the same, but with less deliberation.

Maybe you are correct and that includes all of us on here but I'm not sure it matters that much assuming you save enough and stick to your plan. The thing is none of us know what will work the best in the future. If your plan works reasonably well though (say just average) then you will probably be fine.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #47 on: January 08, 2016, 02:38:58 PM »
Now that gold is "caught up" to inflation, beltim, shouldn't it be able to provide the purpose of it's being in the PP in the first place?

In terms of providing inflation protection over the long term, probably.  But I don't think it will provide any actual real return.  Since 1980, according to two cobbled-together sets of data I pulled up, the CAGR of stocks is 10.7%, while the PP has returned 7.2% over the same period.  And when you consider that this time period has been the best for long-term treasury bonds in history, I would wager a significant amount of money that the long-term return of the PP will be significantly less than 3.5 percentage points less than stocks.

The PP certainly has less volatility than an all-stock portfolio, but I don't think its return will ever be competitive with stocks over the long term again.  I am interested in seeing how it affects SWR, but I think I'd have to go to Monte Carlo simulations in order to properly model it

I think you have to be really careful with the idea that gold is a hedge against inflation. I know that is the common wisdom but I don't think it's true.

https://www.bernstein.com/Content/Research/Publications/WhatsSafe.htm

Over short periods, I totally agree – the price of gold is driven by speculation in the short term.  In the very long term, gold tracks the price of inflation almost better than any other measurement.


Now, whether that's useful to individual investors as a hedge against inflation is a different question.  But I simply said that the price of gold should track inflation over the long term, and I think that's true.

brooklynguy

  • Handlebar Stache
  • *****
  • Posts: 2204
  • Age: 43
Re: In retirement the PP seems to work, what am I missing??
« Reply #48 on: January 08, 2016, 02:53:42 PM »
So can you clarify what you mean by "logic and reason" that are you suggesting gets used to evaluate which portfolio to go with?

I just meant that an investment strategy should never be borne solely out of backtesting, as some are (especially among the various species of market timing), without any good rational explanation to believe that their successful performance will persist into the future.  And, with respect to portfolio construction in particular, at least some of the recommended portfolios out there seem to be the result of curve-fitting from data-mining.  If you can't come up with any good explanation to support an investment strategy (beyond the fact that it worked in the past), then, in my view, it's not a good investment strategy.

This article articulates the above point extremely well:  Philosophical Economics: "Financial Backtesting: A Cautionary Tale"

The well-argued article concludes (among other things) that:

"From an investment perspective, a theoretical understanding of how the market produces a given outcome is important – arguably just as important as 'the data' showing that it does produce that outcome.  We need such an understanding in order to be able to evaluate the robustness of the outcome, the likelihood that the outcome will continue to be seen in the future.  Those that have spent time testing out quantitative approaches in the real world can attest to the fact that the risk that a well-backtested strategy will not work in the future is significant."

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: In retirement the PP seems to work, what am I missing??
« Reply #49 on: January 08, 2016, 03:15:16 PM »
So can you clarify what you mean by "logic and reason" that are you suggesting gets used to evaluate which portfolio to go with?

I just meant that an investment strategy should never be borne solely out of backtesting, as some are (especially among the various species of market timing), without any good rational explanation to believe that their successful performance will persist into the future.  And, with respect to portfolio construction in particular, at least some of the recommended portfolios out there seem to be the result of curve-fitting from data-mining.  If you can't come up with any good explanation to support an investment strategy (beyond the fact that it worked in the past), then, in my view, it's not a good investment strategy.

This article articulates the above point extremely well:  Philosophical Economics: "Financial Backtesting: A Cautionary Tale"

The well-argued article concludes (among other things) that:

"From an investment perspective, a theoretical understanding of how the market produces a given outcome is important – arguably just as important as 'the data' showing that it does produce that outcome.  We need such an understanding in order to be able to evaluate the robustness of the outcome, the likelihood that the outcome will continue to be seen in the future.  Those that have spent time testing out quantitative approaches in the real world can attest to the fact that the risk that a well-backtested strategy will not work in the future is significant."

That is a really, really good article.  Thanks for sharing it.