Rebalancing, another study

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Rebalancing, another study

Post by livesoft »

I came across Alex Wilkinson's web site and a very informative article on portfolio Rebalancing which goes beyond all previous studies that I have ever read about rebalancing. The results are in-line with what has been seen before for the methods that have been published before.

What is MOST INTERESTING are the additional methods that he tested and the take-home message that rebalancing benefits are HIGHLY dependent on the percentage of one's portfolio devoted to equities.

First, a quote:
For portfolios with 50% to 80% invested in stocks, a percent trigger set at 3% or 4% tended to boost returns by 1% to 2% per year, on average, compared to income placement alone. That's a big benefit. The same pattern occurred for all five portfolios in the study.
And next, here is the link: http://www.able2pay.com/rebalancing.html

This should lead to quite a discussion here at Bogleheads. I predict this article will end up as a link in the Bogleheads Wiki. :)
Wiki This signature message sponsored by sscritic: Learn to fish.
mhalley
Posts: 10424
Joined: Tue Nov 20, 2007 5:02 am

Re: Rebalancing, another study

Post by mhalley »

The article seems to vary somewhat from the usual boglehead advice of 5% variance. Quarterly is a little frequent, but not that much at odds with bogleheads. Thanks for the interesting article.
User avatar
jhfenton
Posts: 4754
Joined: Sat Feb 07, 2015 10:17 am
Location: Ohio

Re: Rebalancing, another study

Post by jhfenton »

Thanks for the link. It was an interesting read. I just have a few quibbles related to his using a 133-year time frame.

I do something similar to the cash flow plus simple 3% method that worked so well, with one practical difference.

I keep a cash flow projection by account and individual holding for the current and subsequent calendar year. I manually rebalance when I see that I can't easily get the portfolio back to target using only cash flows during the next 6 months or so. But with our increasing income and annual inflows running 10-12% of the portfolio, that doesn't happen very often. And generally it wouldn't happen very often until late in a normal investor's accumulation phase.

In contrast, in a 133-year-long portfolio test, after a few decades, you'd eventually have so large a portfolio that cash flows would be dwarfed by market volatility. The huge difference he finds from rebalancing versus using income placement and cash flow is not going to be there in a normal human lifespan.

That doesn't mean the recommendation is wrong. It just reduces the magnitude of the benefit from manual rebalancing versus cash flow rebalancing.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

mhalley wrote:The article seems to vary somewhat from the usual boglehead advice of 5% variance. Quarterly is a little frequent, but not that much at odds with bogleheads. Thanks for the interesting article.
The "usual boglehead advice of 5% variance" is something from Larry Swedroe. That is, 5% is not gospel.

The article linked in this thread did NOT conclude that quarterly rebalancing was best or advised.
Last edited by livesoft on Sun Jan 24, 2016 5:11 pm, edited 3 times in total.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

jhfenton wrote:Thanks for the link. It was an interesting read. I just have a few quibbles related to his using a 133-year time frame.
I didn't think the 133-year time frame was important. Other time frames were looked at, too.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
jhfenton
Posts: 4754
Joined: Sat Feb 07, 2015 10:17 am
Location: Ohio

Re: Rebalancing, another study

Post by jhfenton »

livesoft wrote:
jhfenton wrote:Thanks for the link. It was an interesting read. I just have a few quibbles related to his using a 133-year time frame.
I didn't think the 133-year time frame was important. Other time frames were looked at, too.
It matters when considering a cash flow strategy, as the size of the cash flow relative to the portfolio will change over time. After 30 or 40 years, the cash flow growing at roughly inflation is going to be dwarfed by the size of a portfolio with real growth.

Like I said, I don't think it changes the advice, only the magnitude of the benefit of manual rebalancing versus using cash flow alone.

Edited: It's not clear from the little summary charts how they calculated the percentage incremental benefit from each strategy when they tested a variety of portfolios.
longinvest
Posts: 5672
Joined: Sat Aug 11, 2012 8:44 am

Re: Rebalancing, another study

Post by longinvest »

Rebalancing to a fixed stock/bond ratio is a little riskier than the perfect rebalancing method proposed by Nobel Laureate William Sharpe*: select an AA, then adapt the AA to the current market weights of stocks and bonds when rebalancing.

Here's an example:

Initial:
Total market: 60% stocks / 40% bonds
Portfolio AA: 80% stocks / 20% bonds

Rebalance time:
Total market: 48% stocks / 52% bonds:

Adapt the AA to the new market proportions.
- stock change factor: 48%/60% = 0.8
- bond change factor: 52%/40% = 1.3
New AA:
- stocks: 80% x .8 / (80% x .8 + 20% x 1.3) = 71%
- bonds: 20% x 1.3 / (80% x .8 + 20% x 1.3) = 29%

=> Rebalance to 71% stocks / 29% bonds

This makes for smaller rebalancing transactions, and more importantly, if stocks or bonds (but not both) were go to zero, one wouldn't be ruined while continuing to rebalance. As Sharpe explains in his paper, it also avoids contrarian transactions; all market participants could adopt this rebalancing method and the market would still work. Unfortunately, the proportions of bonds and stocks in the market are not easy to find. :annoyed

One shouldn't expect any rebalancing bonus or penalty out of the perfect rebalancing method. It is a market neutral method.

Infrequently rebalancing to a fixed ratio of stocks and bonds is a good enough approximation of the perfect rebalancing method, exposed above. Which trigger is used (fixed date, bands) is not important; it is more important for the rebalancing not to be frequent to minimize market impact.

* William Sharpe paper about Adaptive Asset Allocation Policies: http://web.stanford.edu/~wfsharpe/aaap/wfsaaap.pdf
Last edited by longinvest on Sun Jan 24, 2016 2:40 pm, edited 1 time in total.
Variable Percentage Withdrawal (bogleheads.org/wiki/VPW) | One-Fund Portfolio (bogleheads.org/forum/viewtopic.php?t=287967)
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

@longinvest, I didn't go look up Sharpe's perfect rebalancing method because from your sentences it didn't appear to be perfect. To me a perfect rebalancing method is to rebalance into equities at the low point of equities every year and to rebalance out of equities at the high point before equities drop.
Wiki This signature message sponsored by sscritic: Learn to fish.
longinvest
Posts: 5672
Joined: Sat Aug 11, 2012 8:44 am

Re: Rebalancing, another study

Post by longinvest »

livesoft wrote:@longinvest, I didn't go look up Sharpe's perfect rebalancing method because from your sentences it didn't appear to be perfect. To me a perfect rebalancing method is to rebalance into equities at the low point of equities every year and to rebalance out of equities at the high point before equities drop.
Wouldn't that be awesome!

:sharebeer
Variable Percentage Withdrawal (bogleheads.org/wiki/VPW) | One-Fund Portfolio (bogleheads.org/forum/viewtopic.php?t=287967)
User avatar
LadyGeek
Site Admin
Posts: 95466
Joined: Sat Dec 20, 2008 4:34 pm
Location: Philadelphia
Contact:

Re: Rebalancing, another study

Post by LadyGeek »

livesoft wrote: I predict this article will end up as a link in the Bogleheads Wiki. :)
Put the thread in the wrong forum and it's guaranteed. :)

This thread is now in the Investing - Theory, News & General forum (theory).

I added the article to the wiki: Rebalancing ("External links") which also links back to this thread.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
edge
Posts: 3833
Joined: Mon Feb 19, 2007 6:44 pm
Location: NY

Re: Rebalancing, another study

Post by edge »

This is what I do. I end up rebalancing once or twice every three years.
User avatar
Toons
Posts: 14459
Joined: Fri Nov 21, 2008 9:20 am
Location: Hills of Tennessee

Re: Rebalancing, another study

Post by Toons »

I rebalance gradually.
With New cash,dividends,cap gains ,etc.
I was 90/10 six years ago.
My target is 60/40.
I am 70/30 now.
65 Retired :happy
"One does not accumulate but eliminate. It is not daily increase but daily decrease. The height of cultivation always runs to simplicity" –Bruce Lee
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

LadyGeek wrote:I added the article to the wiki: Rebalancing ("External links") which also links back to this thread.
Much obliged, thanks!
Wiki This signature message sponsored by sscritic: Learn to fish.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Rebalancing, another study

Post by stlutz »

Thanks for the link, livesoft.

I think the three interesting takeaways for me where:

a) I was surprised how closely rebalancing using only dividends kept one to their target allocations over time. Certainly those with large taxable portfolios should take note.

b) I think the study is also consistent with the overall concept that there is shorter term momentum and longer-term mean reversion in returns (or to put it another way, the market doesn't actually fully reflect new/relevant information in prices immediately). I would be skeptical of turning his numbers of exactly what percentage worked best historically into gospel truth going forward. So, I'd still stick with the view that daily rebalancing is unnecessary (harmful?) but one is better off not letting things get too out of whack.

c) There does seem to be value in paying attention to the market. That is, the recent correction probably should prompt one to look at their portfolio to see if rebalancing is warrant as opposed to waiting until some arbitrary date to do it.
User avatar
nedsaid
Posts: 19249
Joined: Fri Nov 23, 2012 11:33 am

Re: Rebalancing, another study

Post by nedsaid »

Okay, okay. I admit it. I have not rebalanced my portfolio after the last downturn. I know, I know. I am a bad example to my fellow Bogleheads. So I suppose I will be lashed with wet noodles on this forum.

My take on rebalancing is that sometimes there is a rebalancing bonus and sometimes there isn't. There is a bonus to the extent you are selling high and buying low. In a long bear market, it would seem to me that rebalancing from bonds to stocks over time would reduce returns for a while. You would keep rebalancing into a falling stock market. It could be quite a while before you saw the benefit from all that rebalancing. Like many other things, a lot probably depends on the timing.
A fool and his money are good for business.
stlutz
Posts: 5585
Joined: Fri Jan 02, 2009 12:08 am

Re: Rebalancing, another study

Post by stlutz »

longinvest--I also read the Sharpe paper. The big problem with it is that one of his premises is simply false. He says it is not possible for all investors to rebalance to a fixed allocation. In reality, it is. If the current market value ratio of stocks/bond is 50/50, and investors collectively decide that they want to be 60/40, the price of stocks *will* rise such that the new ratio of the overall market is now 60/40.

The aggregate market value ratio of stocks to bonds is by definition the value-weighted average of the preference of investors collectively. It incorporates the views of those rebalancing to fixed allocations and those chasing after whatever is "hot".

All Sharpe is doing is looking at the market price and then putting a value judgment on it. Thus he moves from saying that the market ratio reflects investor preferences to saying that the market-based ratio is "correct" and is what all investors should be using.
User avatar
grabiner
Advisory Board
Posts: 35265
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Rebalancing, another study

Post by grabiner »

The study should compare both returns and risk. Failing to rebalance effectively (as with income placement alone, or annual rebalancing in volatile markets) increases the risk of your portfolio; a portfolio which is nominally 50% stock becomes riskier if it sometimes reaches 60% stock. The inadequate rebalancing of income placement hurt investors badly in 1929 and 2000, and cost some of the recovery in 1932 and 2009.

A percentage threshold avoids this issue. If your portfolio is nominally 50% stock, and a rising market takes it to 55% (which would require about a 22% market gain), you go back to 50% and keep the same risk level.

Income placement is still a good idea, particularly in a taxable account, because it allows you to rebalance for free. And you can do the same with allocation of new money. I do both of these in my taxable account, and have never hit the 5% threshold which would require a rebalance there. (I hit the threshold in my portfolio as a whole, selling bonds to buy stocks near the 2008 market bottom, and selling REITs in 2009, but both were in tax-deferred accounts.)
Wiki David Grabiner
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Rebalancing, another study

Post by Rodc »

Interesting.

I think this pretty well lines up with other studies, but has some interesting new ways of slicing the data. I think the benefits claimed are rather larger than most other studies.

I think some others have found similar results. I did a search on Opportunistic Rebalancing but did not find a copy I could access (here is a summary http://www.cfapubs.org/doi/pdf/10.2469/dig.v38.n3.33) This has been discussed here several times.

This paper suggested 20% of target, so for 60% stocks that would be +/-12% which is rather larger. But also showed rather smaller benefits one method to another.

If someone has a link to the paper that works it would be useful. My memory being what it is.

But over all this new study makes sense though I note they forgot to look at RBD rebalancing.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
fortyofforty
Posts: 2083
Joined: Wed Mar 31, 2010 12:33 pm

Re: Rebalancing, another study

Post by fortyofforty »

Very timely, livesoft. I am currently debating, in my mind, whether to begin a periodic rebalancing strategy (simplicity) or continue waiting for the percentage variance (more complicated). This is good reading.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

Rodc wrote:I think some others have found similar results. I did a search on Opportunistic Rebalancing but did not find a copy I could access (here is a summary http://www.cfapubs.org/doi/pdf/10.2469/dig.v38.n3.33) This has been discussed here several times.
This is a llnk in the wiki: http://www.tdainstitutional.com/pdf/Opp ... yanani.pdf
Wiki This signature message sponsored by sscritic: Learn to fish.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Rebalancing, another study

Post by Rodc »

livesoft wrote:
Rodc wrote:I think some others have found similar results. I did a search on Opportunistic Rebalancing but did not find a copy I could access (here is a summary http://www.cfapubs.org/doi/pdf/10.2469/dig.v38.n3.33) This has been discussed here several times.
This is a llnk in the wiki: http://www.tdainstitutional.com/pdf/Opp ... yanani.pdf
Great, thank you. I should have known to look there. :oops:

I note that study was over a much shorter time period.

It would be interesting to know if the larger differences noted in the new study came mostly early in the time period studies or if was seen more or less uniformly over the entire period.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
User avatar
Fat-Tailed Contagion
Posts: 1251
Joined: Fri Mar 02, 2007 10:49 am

Re: Rebalancing, another study

Post by Fat-Tailed Contagion »

What does 3-4% trigger mean exactly ?

Same as the 5% absolute trigger in the 5/25 method ?
“The intelligent investor is a realist who sells to optimists and buys from pessimists.” | ― Benjamin Graham, The Intelligent Investor (75/25 - 50/50 - 25/75)
longinvest
Posts: 5672
Joined: Sat Aug 11, 2012 8:44 am

Re: Rebalancing, another study

Post by longinvest »

OK. Here's a challenge.

An investor has spread his allocation across accounts, as it is often recommended for tax efficiency reasons. Real life being what it is, contributions to various accounts are not symmetric or even regular.

How would this investor rebalance using only distributions and new contributions?

[off topic]
Somehow, a mirror allocation across accounts, while not necessarily most tax-efficient, might be just be good enough, given that it is very difficult to predict future laws and taxes, and that it allows for lazy rebalancing using distributions and contributions only. Maybe it is time to rethink our usual recommendations?
[/off topic]
Last edited by longinvest on Mon Jan 25, 2016 9:07 am, edited 1 time in total.
Variable Percentage Withdrawal (bogleheads.org/wiki/VPW) | One-Fund Portfolio (bogleheads.org/forum/viewtopic.php?t=287967)
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Rebalancing, another study

Post by Rodc »

Fat-Tailed Contagion wrote:What does 3-4% trigger mean exactly ?

Same as the 5% absolute trigger in the 5/25 method ?
3% of "drift"

drift is defined in the article as sum of all positive offsets from targets.

If you want 30% Large US, 20% Small Value US, 30% US TBM and 20% x-US Bonds

and you have 33% Large US, 19% small value, 32% US TBM and 16% x-US bonds the drift is 3% + 2% = 5%.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
longinvest
Posts: 5672
Joined: Sat Aug 11, 2012 8:44 am

Re: Rebalancing, another study

Post by longinvest »

stlutz wrote:longinvest--I also read the Sharpe paper. The big problem with it is that one of his premises is simply false. He says it is not possible for all investors to rebalance to a fixed allocation. In reality, it is. If the current market value ratio of stocks/bond is 50/50, and investors collectively decide that they want to be 60/40, the price of stocks *will* rise such that the new ratio of the overall market is now 60/40.

The aggregate market value ratio of stocks to bonds is by definition the value-weighted average of the preference of investors collectively. It incorporates the views of those rebalancing to fixed allocations and those chasing after whatever is "hot".

All Sharpe is doing is looking at the market price and then putting a value judgment on it. Thus he moves from saying that the market ratio reflects investor preferences to saying that the market-based ratio is "correct" and is what all investors should be using.
Stlutz,

I like your critic. I think that he is trying to get the investor to rebalance with minimal market impact. But, you're right. If everyone wanted to buy more stocks to get to 60/40, prices would adjust appropriately.

Thanks!
Variable Percentage Withdrawal (bogleheads.org/wiki/VPW) | One-Fund Portfolio (bogleheads.org/forum/viewtopic.php?t=287967)
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

Thanks for all the comments everyone.
stlutz wrote:I would be skeptical of turning his numbers of exactly what percentage worked best historically into gospel truth going forward.
I agree that one needn't treat a single trigger number as the best going forward and I didn't get that the author was advocating any specific trigger number though he tested "across triggers of 1%, 2,%, 3%, 4%, 5% and 10%, for monthly, quarterly, and annual rebalancing" for different asset allocations. As folks may have figured out, I have a moving trigger percentage for rebalancing myself. :)

The article/blog does suggest to me some future studies that one might do. I just liked that it got away from some of the rote conventional wisdom espoused on this forum and pointed to other possibilities even while confirming the things most of us already have seen before.
Wiki This signature message sponsored by sscritic: Learn to fish.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Rebalancing, another study

Post by Rodc »

Rodc wrote:
Fat-Tailed Contagion wrote:What does 3-4% trigger mean exactly ?

Same as the 5% absolute trigger in the 5/25 method ?
3% of "drift"

drift is defined in the article as sum of all positive offsets from targets.

If you want 30% Large US, 20% Small Value US, 30% US TBM and 20% x-US Bonds

and you have 33% Large US, 19% small value, 32% US TBM and 16% x-US bonds the drift is 3% + 2% = 5%.
By way of comparison, I use 5/20 bands. That is if US stocks, international stocks or fixed income are off by 5% I rebalance, or if any subasset classes are off by 20% I rebalance. My largest slice is x-US at 20% so that would 4% in this study if it were the only thing high. Other slices are 12.5% or 7.5% so I would trigger at 2.5% and 1.5% absolute. Most likely more than one is above target at any given time, so I think a 3% trigger as defined in this study would trigger more often than what I do.

So these results are not entirely consistent with the other paper linked to above that found wider bands than this to be optimal. Not a direct comparison though because it used % of allocation rather than absolute %.

So, despite this paper claiming higher rewards for rebalancing than other papers (a claim we might be skeptical of until it is better understood just how this arose) I think a meta analysis would suggest that the exact rebalancing method is not terribly important and it is hard to know going forward which method is best.

This method of basing rebalancing on "drift" does have its appeal. I have thought of doing this myself because it takes a portfolio perspective on when you are out of wack vs. individual pieces are out of wack in isolation. That said, if one is hoping to capture a rebalance bonus due to subassets mean reverting perhaps the individual view is better.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
User avatar
Bustoff
Posts: 2033
Joined: Sat Mar 03, 2012 5:45 pm

Re: Rebalancing, another study

Post by Bustoff »

livesoft wrote:...What is MOST INTERESTING are the additional methods that he tested and the take-home message that rebalancing benefits are HIGHLY dependent on the percentage of one's portfolio devoted to equities. ...
Thanks for the link!

Jim Otar wrote an article that appears consistent with the article mentioned by livesoft.
*(Otar uses the presidential election cycle as a yardstick for rebalancing.)

Here are a couple of quotes from the article:
-- "Only at high withdrawal rates and in portfolios with a high equity percentage was annual rebalancing more effective than
rebalancing on the election year only."

-- "There were situations where it was better not to rebalance at all. This happened at low withdrawal rates and in portfolios with a low percentage of equities."
Otar also discusses a different method of rebalancing based on how much your equities grow each year that he calls "growth rebalancing".

Here is a link to the article:
http://retirementoptimizer.com/articles/article47.pdf
User avatar
SpringMan
Posts: 5422
Joined: Wed Mar 21, 2007 11:32 am
Location: Michigan

Re: Rebalancing, another study

Post by SpringMan »

Livesoft,
Thanks for the link.
It has me rethinking my automatically reinvested distributions. I automatically reinvest in tax advantaged accounts and take taxable distributions to MM settlement. It may make good sense to use the distributions toward rebalancing. I have read that a rule of thumb for bond funds is that if you hold them for the duration, you will likely be okay but that assumes dividends are reinvested. I wonder if this is a good reason to automatically reinvest bond fund dividends or not.
Best Wishes, SpringMan
longinvest
Posts: 5672
Joined: Sat Aug 11, 2012 8:44 am

Re: Rebalancing, another study

Post by longinvest »

longinvest wrote: An investor has spread his allocation across accounts, as it is often recommended for tax efficiency reasons. Real life being what it is, contributions to various accounts are not symmetric or even regular.

How would this investor rebalance using only distributions and new contributions?
Bump!
Variable Percentage Withdrawal (bogleheads.org/wiki/VPW) | One-Fund Portfolio (bogleheads.org/forum/viewtopic.php?t=287967)
User avatar
jwillis77373
Posts: 395
Joined: Mon Jun 25, 2007 9:52 pm
Location: Texas

Re: Rebalancing, another study

Post by jwillis77373 »

Interesting that the results are close to compensating for Inflation.
User avatar
grabiner
Advisory Board
Posts: 35265
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Rebalancing, another study

Post by grabiner »

longinvest wrote:OK. Here's a challenge.

An investor has spread his allocation across accounts, as it is often recommended for tax efficiency reasons. Real life being what it is, contributions to various accounts are not symmetric or even regular.

How would this investor rebalance using only distributions and new contributions?
In this situation, it wouldn't work unless the investor holds bonds in both accounts. (And holding bonds in both accounts is often wrong for reasons other than tax efficiency; some retirement plans, such as the TSP and TIAA-CREF, have unusually good fixed income options.) An investor with only stock in taxable can keep the taxable account balanced among different stock categories with distributions and new contributions, but will have less free money to keep the balance between stocks and bonds.

This is my situation. My taxable account is all stock. I invest dividends and new contributions into the most underweighted stock, so I keep my balance between US and foreign stock. However, I cannot divert my stock distributions to the bond fund in my retirement account, and in rising stock markets, I don't always have enough new investment to direct to the bond fund, so I also need to rebalance annually and with a trigger.
Wiki David Grabiner
magneto
Posts: 1060
Joined: Sun Dec 19, 2010 9:57 am
Location: On Chesil Beach

Re: Rebalancing, another study

Post by magneto »

Just love these posts about rebalancing :happy
When will we achieve the ultimate rebalancing method. :?:

Where the article may fall down is failing to take full note of momentum, deviation and time; and hence the extent of rebalancing .
Taking parallels with Classical Control Theory - Integral Action, where system stability is paramount, (the investor being part of a control system), we would be guided to a rebalancing response equivalent to the area under the graph (deviaton * time).

The nearest anyone has come to this Integral Action to date IMHO is from our fellow poster 'Beliavsky', with his 10% deviation rebalance approach.

It should come as no surprise that a 50/50 volatile/non-volatile mix will maximise any rebalancing bonus over a market cycle,. But an investor who is considering just two asset classes such as stocks/bonds, might note the long-term outperformance (the trend) of stocks, and compromise by tilting to stocks, to somewhere in the traditional region of 60/40 to 80/20.

If some account is to be taken of that irrational wretched fellow ‘momentum’ (aka the band-waggon effect), the investor can approach any target at a controlled rate, say 10% of observed deviation from target per month, keeping 90% back each month to benefit from 'momentum' (suggest this method now be referred to as Beliavsky's Rule)..

Good Luck to us All
'There is a tide in the affairs of men ...', Brutus (Market Timer)
sadie wess
Posts: 320
Joined: Wed Jan 19, 2011 8:42 am

Re: Rebalancing, another study

Post by sadie wess »

nedsaid wrote:Okay, okay. I admit it. I have not rebalanced my portfolio after the last downturn. I know, I know. I am a bad example to my fellow Bogleheads. So I suppose I will be lashed with wet noodles on this forum.

My take on rebalancing is that sometimes there is a rebalancing bonus and sometimes there isn't. There is a bonus to the extent you are selling high and buying low. In a long bear market, it would seem to me that rebalancing from bonds to stocks over time would reduce returns for a while. You would keep rebalancing into a falling stock market. It could be quite a while before you saw the benefit from all that rebalancing. Like many other things, a lot probably depends on the timing.
Nedsaid,

Thank you for helping me out today. I did not get to my rebalancing last year or yet this year. I have been overwhelmed with other responsibilities and I am still snowed in and cannot get out of my house. I worry about not having completed my "financial stuff". Your post gives me a little sense of relief. I appreciate that.

I am not the only to get lashed with wet noodles!

:sharebeer

My best regards,

Sadie
User avatar
TomP10
Posts: 238
Joined: Tue Mar 18, 2014 4:13 am

Re: Rebalancing, another study

Post by TomP10 »

I would love to be able to read the entire Swedroe article cited in wiki -- but I can only read the first two pages.

Anyone have any suggestions on browser settings so I can see the entire article?

Tom
"It is remarkable how much long term advantage people like us have gotten by trying to be consistently not stupid, instead of trying to be very intelligent." -- Charlie Munger
rkhusky
Posts: 17652
Joined: Thu Aug 18, 2011 8:09 pm

Re: Rebalancing, another study

Post by rkhusky »

I note that the author found that checking weekly was better than less frequently and referenced a study that found that daily was better than weekly. I presume that in this study the rebalancing occurred on the next trading day at the next day's prices. If one had an automatic rebalancing capability and used ETF's, they could check every second and perform multiple rebalances during a day.
User avatar
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: Rebalancing, another study

Post by siamond »

TomP10 wrote:I would love to be able to read the entire Swedroe article cited in wiki -- but I can only read the first two pages.

Anyone have any suggestions on browser settings so I can see the entire article?
The wiki's advice is: (Disable cookies to view on a single page. Otherwise, a subscription will be required to view.)

Personally, I did subscribe, this is free, and removes this kind of annoyance whenever I check an article or another.
longinvest
Posts: 5672
Joined: Sat Aug 11, 2012 8:44 am

Re: Rebalancing, another study

Post by longinvest »

rkhusky wrote:I note that the author found that checking weekly was better than less frequently and referenced a study that found that daily was better than weekly. I presume that in this study the rebalancing occurred on the next trading day at the next day's prices. If one had an automatic rebalancing capability and used ETF's, they could check every second and perform multiple rebalances during a day.
Yet, the same returns could probably be achieved by simply increasing the stock allocation by a few percents and rebalancing once a year without bands. I think that we're counting the angels on the head of a pin.

I am just of the opinion that how one rebalances is not as important as making sure one will stick to his chosen plan. If one prefers to use rebalancing bands, then he should go with them. If one only wants to rebalance into one direction (stocks to bonds), that's fine. If one does not want to rebalance and wishes to let the market rebalance for him like illustrated in the linked article for the 1920s and 2000s, that's fine too. All these methods (and others) are fine as long as the investor is able to stay the course.

I continue to think that rebalancing is done for managing risk, not to increase returns. As long as equities deliver higher returns, any method that lets the stock allocation stay higher than AA target for significant portions of time is likely to deliver higher portfolio returns.
Variable Percentage Withdrawal (bogleheads.org/wiki/VPW) | One-Fund Portfolio (bogleheads.org/forum/viewtopic.php?t=287967)
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

longinvest wrote:I continue to think that rebalancing is done for managing risk, not to increase returns. As long as equities deliver higher returns, any method that lets the stock allocation stay higher than AA target for significant portions of time is likely to deliver higher portfolio returns.
I certainly can agree with those statements. And those statements always get stated whenever a significant rebalancing bonus shows up in a paper or a study. That is, someone comes along and says, "Oh, of course they did better because they had a higher allocation to stocks! Duh! One has to show me the risk-adjusted rebalancing bonus." I am guilty of saying these things myself.

I would like to suggest that perhaps stocks do NOT have the same risk level at all times. Sometimes stocks are more risky than usual and sometimes stocks are less risky than usual. For instance, if on Friday, January 22 large-cap foreign stock index funds go up by more than 3%, then before Monday January 25, those stocks are riskier than they were on Thursday January 21 when large-cap foreign stock index ETFs had dropped more than 3% in the morning. But then folks will say, "But that's just market timing!" So what if it is as long as it works! And before one jumps on this and says "But market timing never works!" Prove that it doesn't work in this case by some back-testing. And for those that say, "Do your own back-testing!" what if I said, I did, and this market timing does work. ?
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

Bustoff wrote:Jim Otar wrote an article that appears consistent with the article mentioned by livesoft.
*(Otar uses the presidential election cycle as a yardstick for rebalancing.)
I have always been skeptical of the Otar presidential election cycle because of the way the back testing was done. If back testing happens to time something like 1929 or 2009 perfectly, then just a few data points will swamp any other things going on. Perhaps a way of changing the backtesting is to leave out some of these dates and see if the affect is of statistical significance. You know, cross validation, out-of-sample testing, etc.
Wiki This signature message sponsored by sscritic: Learn to fish.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Rebalancing, another study

Post by Rodc »

livesoft wrote:
Bustoff wrote:Jim Otar wrote an article that appears consistent with the article mentioned by livesoft.
*(Otar uses the presidential election cycle as a yardstick for rebalancing.)
I have always been skeptical of the Otar presidential election cycle because of the way the back testing was done. If back testing happens to time something like 1929 or 2009 perfectly, then just a few data points will swamp any other things going on. Perhaps a way of changing the backtesting is to leave out some of these dates and see if the affect is of statistical significance. You know, cross validation, out-of-sample testing, etc.
That is a good point, 2000 was an election year.

I once reverse engineered a few studies back around 2002 or so, and what happened was whatever randomly happened to keep you in stock up close to the top of the tech bubble was the winner because it swamped everything that came before, but I did not see any reason to believe the winner would win if things had turned out a little differently.

Kind of like when I did the RBD study. Regular bands were a little better over the data I had, but as you watched day by day with the two side by side you could see there were periods where one was better than the other, but they the tide turned, and turned again, etc. and so hard to really say regular bands should be expected to win going forward.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
User avatar
TomP10
Posts: 238
Joined: Tue Mar 18, 2014 4:13 am

Re: Rebalancing, another study

Post by TomP10 »

TomP10 wrote:
I would love to be able to read the entire Swedroe article cited in wiki -- but I can only read the first two pages.

Anyone have any suggestions on browser settings so I can see the entire article?
The wiki's advice is: (Disable cookies to view on a single page. Otherwise, a subscription will be required to view.)

Personally, I did subscribe, this is free, and removes this kind of annoyance whenever I check an article or another.
I followed those directions and it did not work. Any other suggestions would be appreciated.

Tom
"It is remarkable how much long term advantage people like us have gotten by trying to be consistently not stupid, instead of trying to be very intelligent." -- Charlie Munger
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

Rodc wrote:... and so hard to really say regular bands should be expected to win going forward.
I don't like regular rebalancing bands either.
Wiki This signature message sponsored by sscritic: Learn to fish.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Rebalancing, another study

Post by Rodc »

livesoft wrote:
Rodc wrote:... and so hard to really say regular bands should be expected to win going forward.
I don't like regular rebalancing bands either.
The one study we have they beat RBD. :)

(though not by much)
(I agree with longinvest above that it is likely wasted energy too worry much about the details of precisely how one rebalances. One should pick any reasonably rational choice and be done with it.)
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
User avatar
XdUzHa3NtSeIkBkIGPVn
Posts: 46
Joined: Thu Jul 16, 2015 4:12 pm

Re: Rebalancing, another study

Post by XdUzHa3NtSeIkBkIGPVn »

longinvest wrote:
stlutz wrote:longinvest--I also read the Sharpe paper. The big problem with it is that one of his premises is simply false. He says it is not possible for all investors to rebalance to a fixed allocation. In reality, it is. If the current market value ratio of stocks/bond is 50/50, and investors collectively decide that they want to be 60/40, the price of stocks *will* rise such that the new ratio of the overall market is now 60/40.

The aggregate market value ratio of stocks to bonds is by definition the value-weighted average of the preference of investors collectively. It incorporates the views of those rebalancing to fixed allocations and those chasing after whatever is "hot".

All Sharpe is doing is looking at the market price and then putting a value judgment on it. Thus he moves from saying that the market ratio reflects investor preferences to saying that the market-based ratio is "correct" and is what all investors should be using.
Stlutz,

I like your critic. I think that he is trying to get the investor to rebalance with minimal market impact. But, you're right. If everyone wanted to buy more stocks to get to 60/40, prices would adjust appropriately.

Thanks!
I'm not sure about this criticism. Sharpe wasn't talking about an individual investor changing his asset allocation at all. The "traditional" allocation is fixed, while the market weights will fluctuate.

So in your example, this would be the market weights changing to 60/40 while individual investors all hold on to the 50/50 allocation they chose years before. If everyone did this, there would be no one to sell their stocks to, and no one to buy bonds from. Sharpe's point is that not everyone in the market could use this contrarian approach, or else there wouldn't be anyone at the other side of rebalancing trades.

What drives it home to me though, is that selling stocks merely because they've gone up in price, or buying merely because because they've gone down implies the market is not efficiently pricing the prospects of future risk-adjusted returns across asset classes. If you don't believe in active management within an asset class, there's no reason to believe in it across asset classes.

I think some adjustments can be made on top of the system described by Sharpe, but to me it has become the starting point for my allocation, as opposed to an absolute percentage of stocks and bonds irrespective of market weights.
User avatar
nedsaid
Posts: 19249
Joined: Fri Nov 23, 2012 11:33 am

Re: Rebalancing, another study

Post by nedsaid »

sadie wess wrote:
nedsaid wrote:Okay, okay. I admit it. I have not rebalanced my portfolio after the last downturn. I know, I know. I am a bad example to my fellow Bogleheads. So I suppose I will be lashed with wet noodles on this forum.

My take on rebalancing is that sometimes there is a rebalancing bonus and sometimes there isn't. There is a bonus to the extent you are selling high and buying low. In a long bear market, it would seem to me that rebalancing from bonds to stocks over time would reduce returns for a while. You would keep rebalancing into a falling stock market. It could be quite a while before you saw the benefit from all that rebalancing. Like many other things, a lot probably depends on the timing.
Nedsaid,

Thank you for helping me out today. I did not get to my rebalancing last year or yet this year. I have been overwhelmed with other responsibilities and I am still snowed in and cannot get out of my house. I worry about not having completed my "financial stuff". Your post gives me a little sense of relief. I appreciate that.

I am not the only to get lashed with wet noodles!

:sharebeer

My best regards,

Sadie
Rebalancing is the next thing to Godliness to some on this forum. I have seen John Bogle interviewed on this subject and he seems to have a rather relaxed attitude about rebalancing.

If you just think about it, if you do nothing, you portfolio will grow more and more stock heavy since stocks have a better return than bonds or cash over time. So if you never rebalance, your returns over time would be a bit higher.

Pretty much, I rebalance when absolutely necessary and that is about it.

I chuckled through your post. Thanks for your comments.

Best wishes,

Ned
A fool and his money are good for business.
User avatar
telemark
Posts: 3389
Joined: Sat Aug 11, 2012 6:35 am

Re: Rebalancing, another study

Post by telemark »

longinvest wrote:
longinvest wrote: An investor has spread his allocation across accounts, as it is often recommended for tax efficiency reasons. Real life being what it is, contributions to various accounts are not symmetric or even regular.

How would this investor rebalance using only distributions and new contributions?
Bump!
I've been thinking about this recently, since I like to rebalance with new money but I just rearranged my two 401Ks so that the international and REIT allocations are all in the old plan, and these are both volatile enough to benefit from new money at fairly regular intervals. One option is just to buy them in the new plan when needed--there are funds for that in the new plan, I just don't like them very much--and occasionally clean house as needed. It's not clear if this would accomplish anything, other than keeping me amused.

The other option is just to add money to the two funds in the new plan, and do an overall rebalance when things get too far out of whack.
User avatar
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: Rebalancing, another study

Post by siamond »

I finally took time to read this new article, and it seems reasonably well grounded. I am also puzzled by the size of the 'bonus' the authors obtained, but their general strategy makes sense, and I remember reading another article reaching similar conclusions (wait for a sizable 'drift' from your regular AA, watch it often, act decisively). Now if we push this logic a bit farther, if one were to monitor daily, while staying disciplined about the sizable drift, one should do even better... Which I think was demonstrated by the other paper (sorry, forgot the reference). While the basic annual rebalancing has little chance of displaying much of a bonus.

Something really cool is that, with minimal efforts using a Google spreadsheet, one can check in seconds what's the status of a given AA, and the drift of one's portfolio. The only manual thing I couldn't automate is that, from time to time, I need to update the # of shares of my various investments (notably if you reinvest dividends). Then it's just one click to see the drift. And the whole thing makes MUCH MORE REALISTIC the use of frequent monitoring. As long as you remember to NOT frequently act, which is the hard part!
Topic Author
livesoft
Posts: 85971
Joined: Thu Mar 01, 2007 7:00 pm

Re: Rebalancing, another study

Post by livesoft »

One doesn't need to frequently look at a spreadsheet to know if one is approaching a drift trigger. One can use spreadsheet to predict what price movement would cause the trigger. Then set that as an alert to your smart phone.

Most days one would be no where near any trigger point. One will know the night before if one is getting close to a trigger point and then set up the alert. One would probably only need to set the alert maybe 1 to 3 times a year.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
siamond
Posts: 6003
Joined: Mon May 28, 2012 5:50 am

Re: Rebalancing, another study

Post by siamond »

livesoft wrote:One doesn't need to frequently look at a spreadsheet to know if one is approaching a drift trigger. One can use spreadsheet to predict what price movement would cause the trigger. Then set that as an alert to your smart phone.
Aaaah... Fancy... Will investigate, this is cool. Thank you.
Post Reply