Matter-scatter, part 4: Bond allocation

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42189
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Matter-scatter, part 4: Bond allocation

Post by nisiprius »

Part 1: Introduction, Part 2: Bond fund choices,, Part 3, International bonds. Brief notes: 1) In these Monte Carlo simulations in which the return for a given month is varied by randomly choosing the actual historical return or the actual historic return for an adjacent month. 2) They show the final value from $100 monthly contributions made over the whole time period. 4) Green and red crosses mark actual historical values. 5) Green and red bars show the 10% percentile, median, and 90% percentile of the range of outcomes. 5) The actual data source is PortfolioVisualizer.com, Backtest Portfolio, Monthly Returns; this is used as calculation input; and no PV content or numeric values are directly reproduced.

Having looks at two cases where changes in the choice of bond asset class didn't matter a great deal, you might wonder if anything matters a great deal, or whether my methodology is a view from so high up that nothing will ever matter.

Bond allocation, that is to say the choice of stock/bond percentages, turns out to matter a lot, although as always there are debates over what any of it means.

The most basic comparison is, simply, between a bond fund and a stock fund:

Image

It is the only chart in this series of postings we will see in which the 10%-90% percentile ranges do not overlap. Even so, you can see visually in the scatter plots that the minimum stock outcome is actually lower than the minimum bond outcome.

The next three comparisons show comparisons between the four Vanguard LifeStrategy funds. Their composition has varied somewhat over time, but in they have stayed close to their present overall stock-bond allocations:

VASIX, Vanguard LifeStrategy Income: 20% stocks, 80% bonds
VSCGX, Vanguard LifeStrategy Conservative Growth: 40% stocks, 60% bonds
VSMGX, Vanguard LifeStrategy Moderate Growth, 60% stocks, 40% bonds
VASGX, Vanguard LifeStrategy Growth, 80% stocks, 20% bonds.

As you can see, each of the four funds is distinctly different from the others.

Image
Image
Image

Our final comparison involves the difference between 100% stocks and 90% stocks, 10% bonds. Vanguard's target retirement dates use about a 90/10 allocation for young investors, and it is often asked whether the 10% bond allocation makes any important difference.

Image

Observations: Notice that in every case, changes in bond allocation produce very visible difference in the standard deviation of the resulting portfolio. This may or may not seem important, depending on your personal feelings about risk. John C. Bogle has written that "An extra percentage point of standard deviation is meaningless, but an extra percentage point of return is priceless." However, what my exploration in this series shows is that a) an extra percentage point of return is hard to come by; b) an change of 1% in standard deviation merely by changing stock portfolio composition is also very hard to come by; c) much larger changes in standard deviation are easily obtained through changes in bond percentage allocation. Bonds seem to move the needle on risk; stock portfolio choices do not.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
Munir
Posts: 2754
Joined: Mon Feb 26, 2007 4:39 pm
Location: Oregon

Re: Matter-scatter, part 4: Bond allocation

Post by Munir »

nisiprius,
Can you please describe and analyze your graphs and conclusions in "plain" language that novices like me can understand?
Thank you.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42189
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Matter-scatter, part 4: Bond allocation

Post by nisiprius »

Munir wrote: Tue Jul 31, 2018 6:33 pm nisiprius,
Can you please describe and analyze your graphs and conclusions in "plain" language that novices like me can understand?
Thank you.
I am somewhat trying to avoid conclusions, because different people are going to draw different conclusions from the same data, but, overall, the things I get from my exploration are: first, the systematic differences resulting from of the kinds of changes we tend to discuss in the forum--is it better to use Total Bond or all Treasuries, say--are very small compared to the natural uncertainty of any portfolio that has a "normal" stock allocation. The part that's within our control, through portfolio selection, is small compared to the part that's out of our control. You then have to add to that the problem that "control" just means "we can change it," but the arguments come because we don't know which way it will go!

This part, part 4, is rather an exception to the others, because unlike changes within the stock portion or the bond portion, the overall balance of stocks to bonds does make a big difference. But that doesn't say which is preferable!

Image

Looking at this specific example, the green parts represent 100% stocks and the red parts represent 90% stocks, 10% bonds. The chart is saying that historically, the investor with 100% stocks ended up with $120K, $5,000 more than the 90/10 investor. The chart is also saying that in a Monte Carlo simulation, realistically, the 100% stocks investor might well have ended up with as little as $77K or as much as $195K. So, we are talking about a $5K difference, within a $118K spread.

We also see, looking at the horizontal axis, that the 90/10 investor did reduce the volatility of the portfolio from maybe 14.5% to 12.5%. Which is not going to be very noticeable in real life.

Is it "worth" the extra volatility of 100% stocks (green) to get the results shown in the green bar rather than the red bar? Who knows? The big point is that you should not be thinking of it as if there were perfect certainty that you'd get $5,000 more. It is $5,000 more in the context of $118K uncertainty about how much you will get.

From the introduction, linked in the fine print:
Here's what the chart is showing.
  • The green and red dots are the final outcomes of 500 Monte Carlo simulations of the possible performance of the two funds.
  • The performance is not based on the growth of a single investment in the fund. It is, instead, based on the outcome of making periodic fund purchases of $100/month over the time period shown. It thus is closer to the way most of us actually invest.
  • The green and red crosses mark the outcome based on the actual historical performance of the funds.
  • The red and green range lines show the 10% and 90% percentiles of those results. In this case, 80% of the simulations showed final values for VFINX between $72,000 and $182,000; 10% were below $72,000, 10% were above $182,000. The middle mark, $113K, is the median.
  • The basis for all of the data that went into the simulation is the set pf "monthly values" tabulated by PortfolioVisualizer, "Backtest Portfolio" tool.
  • The range of time shown is that for all available data in PortfolioVisualizer.
  • The vertical scale represents the final value in dollars.
  • The horizontal scale is the annualized standard deviation of the monthly returns of the plotted assets. It is a measure of how much fluctuation you would see in portfolio value.
Last edited by nisiprius on Wed Aug 01, 2018 8:36 am, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
onthecusp
Posts: 817
Joined: Mon Aug 29, 2016 3:25 pm

Re: Matter-scatter, part 4: Bond allocation

Post by onthecusp »

nisiprius wrote: Wed Aug 01, 2018 8:16 am
The big point is that you should be thinking of it as if there were perfect certainty that you'd get $5,000 more. It is $5,000 more in the context of $118K uncertainty about how much you will get.
Great stuff nisprius.

I was wondering about your big point. Is there a typo? I would think that there were NOT perfect certainty.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 42189
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Matter-scatter, part 4: Bond allocation

Post by nisiprius »

onthecusp wrote: Wed Aug 01, 2018 8:34 am
nisiprius wrote: Wed Aug 01, 2018 8:16 am
The big point is that you should be thinking of it as if there were perfect certainty that you'd get $5,000 more. It is $5,000 more in the context of $118K uncertainty about how much you will get.
Great stuff nisprius.

I was wondering about your big point. Is there a typo? I would think that there were NOT perfect certainty.
:oops:
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
GAAP
Posts: 1154
Joined: Fri Apr 08, 2016 12:41 pm

Re: Matter-scatter, part 4: Bond allocation

Post by GAAP »

nisiprius wrote: Wed Aug 01, 2018 8:16 am The part that's within our control, through portfolio selection, is small compared to the part that's out of our control. You then have to add to that the problem that "control" just means "we can change it," but the arguments come because we don't know which way it will go!
Nice synopsis -- something we all need to remember.

Another thing I take from these charts is that quoting a specific historical result is essentially the same as randomly picking one of those dots on your charts -- there's a reason companies use those forward looking statement disclosures...
User avatar
CaliJim
Posts: 3050
Joined: Sun Feb 28, 2010 8:47 pm
Location: California, near the beach

Re: Matter-scatter, part 4: Bond allocation

Post by CaliJim »

So... you are reconfirming that top level stock to bond allocation matters a lot. Well done! Thanks.

Your charts do not make clear what happens during draw down, and some folks might come to the wrong conclusion. People sometimes overestimate their risk tolerance based on the historically better long term performance of stocks over bonds.

Would you be willing to simulate and chart what happens if you now run the MC Sim in reverse? What would it look like starting with, say, 200k in various stock vs bond allocations, withdrawing $100 a month instead of contributing, and scatter plotting the final outcomes and sigmas?

The reason I ask is that, stock performance being so much 'better', folks might be tempted to go "Stock for 100%, Mr Trebek"..

IIRC, the simulations for 100% stock portfolios ain't so hot during retirement drawdown....
Last edited by CaliJim on Thu Aug 02, 2018 10:53 pm, edited 1 time in total.
-calijim- | | For more info, click this Wiki
User avatar
onthecusp
Posts: 817
Joined: Mon Aug 29, 2016 3:25 pm

Re: Matter-scatter, part 4: Bond allocation

Post by onthecusp »

nisiprius wrote: Wed Aug 01, 2018 8:37 am
onthecusp wrote: Wed Aug 01, 2018 8:34 am
nisiprius wrote: Wed Aug 01, 2018 8:16 am
The big point is that you should be thinking of it as if there were perfect certainty that you'd get $5,000 more. It is $5,000 more in the context of $118K uncertainty about how much you will get.
Great stuff nisprius.

I was wondering about your big point. Is there a typo? I would think that there were NOT perfect certainty.
:oops:
So hard to proof my own writing. You do wonderfully with yours and you are willing to put a lot out there. :sharebeer

On the subject of "certainty", I'm looking at the big crosses in the middle of each cloud that represents the actual performance of the fund or portfolio versus the little dots that represent individual results from each monte carlo run.

I'm seeing that all the big crosses in these scatter plots are in fact very close to the mean of the smaller dots in both axis. If one small dot is about as likely to be the actual path as another, shouldn't there be a few funds that show a significant offset from the middle of the cloud? Just as a matter of random chance?
Bob
Posts: 211
Joined: Mon Feb 26, 2007 5:15 pm

Re: Matter-scatter, part 4: Bond allocation

Post by Bob »

Great analysis and much appreciate that you have shared it with us all.

1+ CaliJim's suggestion (request?) that this would be a great analysis to see for a decumulation scenario. Specifically for the bond allocation aspect.
boglesmind
Posts: 235
Joined: Sun Jan 05, 2014 1:07 pm

Re: Matter-scatter, part 4: Bond allocation

Post by boglesmind »

+ Infinity! to the 4-part series.

Boglesmind
Post Reply