Dynamic AA - does it make sense?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.

Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 10:50 am

Hi,

I have been pondering about the following. Say you have a standard AA between stocks and bonds, for example 70/30. Does it, in your opinion, ever make sense to forgo your 70/30, and go to 90/10 (or 80/20 or 100/0 or ...) during a certain period because of bad past results of the equities? If you had 70/30 before 2008, and went to 100/0 (the extreme) after/during the market collapsed, you would be buying low. Years later, you will benefit.

To begin with, I don't think this is market timing, since you are over weighting assets which have done bad in the past.

You need some kind of threshold to decide that you forgo your AA in favor of more equities. A threshold might be that the market has been down for XX% (for example 30%) since it's highs. You can never predict (and you don't want to try - market timing) the lowest point, so you need some kind of threshold. Once the equities have been recovered - again, a threshold is needed here -, you can again convert to 70/30.

Does this make sense? What is your opinion?
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 11:11 am

If you had 70/30 before 2008, and went to 100/0 (the extreme) after/during the market collapsed, you would be buying low. Years later, you will benefit.

To begin with, I don't think this is market timing, since you are over weighting assets which have done bad in the past.


Of course it is blatant market timing.

The problem is that 2008 could have been much worse. Unheard of intervention by the Fed and federal agencies might not have been taken for one thing. In the past these sorts of market collapses dropped P/E10 below 10. If you waited for that you never made the switch. If you did make the switch at above 10, history shows the market can drop much lower. Whoops.

And often when these things happen you end up unemployed and you need to sell stocks to eat. You buy at P/E10 of say 12, you then lose a bunch from there, and end up selling at P/E10 of 8 to eat and pay rent or mortgage. Then your investment get slaughtered.

So, sure, this might make you a mint. It might be a disaster. Just like you might expect from a risky move. It certainly is not a slam dunk winning strategy.

In fact it is often said in the great depression that all the stupid folks lost their money in the first crash and all the smart money got lost in the second crash a few year later (for just this sort of reason).

I don't think the suggestion is crazy. I do think you underestimate the dangers and over estimate the odds of success.
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby magneto » Tue Jul 16, 2013 11:22 am

I think the figures quoted are a little extreme. A number of posters in previous threads have suggested altering allocations (over-balancing) based on stock valuations such as PE10.

Prefer to set a central AA from which I can vary a little depending on the attractiveness of the various asset classes. Most of the time it is far from clear whether stocks are over or under valued and a central AA is held. Occasionally such as 1999, 2008, a clear picture does emerge that stocks are moving to extremes and IMHO over-balancing can then be productive. In moderation.
Last edited by magneto on Tue Jul 16, 2013 11:28 am, edited 1 time in total.
'There is a tide in the affairs of men ...', Brutus (Market Timer)
magneto
 
Posts: 299
Joined: Sun Dec 19, 2010 11:57 am

Re: Dynamic AA - does it make sense?

Postby YDNAL » Tue Jul 16, 2013 11:25 am

AppelSienSapFrietjes wrote:I have been pondering about the following. Say you have a standard AA between stocks and bonds, for example 70/30. Does it, in your opinion, ever make sense to forgo your 70/30, and go to 90/10 (or 80/20 or 100/0 or ...) during a certain period because of bad past results of the equities? If you had 70/30 before 2008, and went to 100/0 (the extreme) after/during the market collapsed, you would be buying low. Years later, you will benefit.

How do you know when the market stops "collapsing" ? How do you know when to buy for "years later to benefit" ?
  • We can look back and say... there, we should go 100% stocks around August 2003 and March 9, 2009.
  • Try to do that going forward !! - consistent market timing doesn't work (you may get lucky here/there).

Image
Landy | Be yourself, everyone else is already taken -- Oscar Wilde
YDNAL
 
Posts: 13404
Joined: Tue Apr 10, 2007 5:04 pm
Location: Biscayne Bay

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 11:32 am

YDNAL wrote:
AppelSienSapFrietjes wrote:I have been pondering about the following. Say you have a standard AA between stocks and bonds, for example 70/30. Does it, in your opinion, ever make sense to forgo your 70/30, and go to 90/10 (or 80/20 or 100/0 or ...) during a certain period because of bad past results of the equities? If you had 70/30 before 2008, and went to 100/0 (the extreme) after/during the market collapsed, you would be buying low. Years later, you will benefit.

How do you know when the market stops "collapsing" ? How do you know when to buy for "years later to benefit" ?
  • We can look back and say... there, we should go 100% stocks around August 2003 and March 9, 2009.
  • Try to do that going forward !! - consistent market timing doesn't work (you may get lucky here/there).

Image


You don't know when the markets stop collapsing. That's not the point. You have a certain threshold (for example -30%), and then you alter in favor of equities. You might hit the gold spot (the very low), or not, you don't know at that point in time. But that's not really important. You are still better by this, than you when you keep your AA.

Whether you are 100/0, 90/10 or 60/40, you still believe in the markets. If not, you would have been (or would have wanted to go) to 0/100 or so. Given that assumption, confidence, it isn't a bad thing to alter your AA in favor of equities at certain times?

Once the market has been stabilized, you can again change to your original AA.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Chris M » Tue Jul 16, 2013 11:48 am

Rodc wrote:
If you had 70/30 before 2008, and went to 100/0 (the extreme) after/during the market collapsed, you would be buying low. Years later, you will benefit.

To begin with, I don't think this is market timing, since you are over weighting assets which have done bad in the past.


Of course it is blatant market timing.


I don't think it's market timing as long as you recognize that stocks could continue to go down. In that case you are simply buying low in the expectation of an eventual reward, without trying to time the bottom. I think you could view it as being similar to rebalancing--in effect you would be "over-rebalancing." I rebalanced out of bonds and into stocks in January 2009 not because I wanted to (it made me nervous) or because I thought the market had bottomed, but because I always rebalance in January. I expected the market to continue down--and sure enough it did for another 2 months.

It's a question I've been wondering about myself. If there is a rebalancing bonus, then it seems that the bonus would be even bigger if you "over-rebalance", no?
Chris M
 
Posts: 88
Joined: Tue May 28, 2013 7:25 pm

Re: Dynamic AA - does it make sense?

Postby YDNAL » Tue Jul 16, 2013 12:02 pm

AppelSienSapFrietjes wrote:You don't know when the markets stop collapsing. That's not the point. You have a certain threshold (for example -30%), and then you alter in favor of equities. You might hit the gold spot (the very low), or not, you don't know at that point in time. But that's not really important. You are still better by this, than you when you keep your AA.

Whether you are 100/0, 90/10 or 60/40, you still believe in the markets. If not, you would have been (or would have wanted to go) to 0/100 or so. Given that assumption, confidence, it isn't a bad thing to alter your AA in favor of equities at certain times?

Once the market has been stabilized, you can again change to your original AA.

You can market-time all you want, AppelSienSapFrietjes, it is YOUR money. I just gave you food for thought.
Landy | Be yourself, everyone else is already taken -- Oscar Wilde
YDNAL
 
Posts: 13404
Joined: Tue Apr 10, 2007 5:04 pm
Location: Biscayne Bay

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 12:12 pm

YDNAL wrote:
AppelSienSapFrietjes wrote:You don't know when the markets stop collapsing. That's not the point. You have a certain threshold (for example -30%), and then you alter in favor of equities. You might hit the gold spot (the very low), or not, you don't know at that point in time. But that's not really important. You are still better by this, than you when you keep your AA.

Whether you are 100/0, 90/10 or 60/40, you still believe in the markets. If not, you would have been (or would have wanted to go) to 0/100 or so. Given that assumption, confidence, it isn't a bad thing to alter your AA in favor of equities at certain times?

Once the market has been stabilized, you can again change to your original AA.

You can market-time all you want, AppelSienSapFrietjes, it is YOUR money. I just gave you food for thought.


Sorry, don't get me wrong, I appreciate your feedback and food for thought! English is not my native language, so forgive me if I am unable to make myself clear all the time and sound harsh now and then.

A few threads above and years ago, Rick Ferri was shouting in 2008 that young people should go all-in for equities ("Early Savers (20s and 30s) - buy equities index funds like crazy with what you can and do not look at your account balance for 10 years."). W. Buffet was making the same remarks in 2008.

Why isn't that timing? I fail to see the difference between their proposal and my proposal.

Is a safe conclusion that this is a valid option during wealth accumulation (in other words, for younger people - 20s and 30s), and a bad idea for older people?
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 12:22 pm

If there is a rebalancing bonus, then it seems that the bonus would be even bigger if you "over-rebalance", no?


There is no consistent rebalancing bonus. Many threads on the subject so won't explain here, other than to point out that over time most rebalancing is from high growth to low growth so that high growth does not become too high an allocation; ie from stocks to bonds which lowers returns (and risk which is the true point of 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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 12:34 pm

Why isn't that timing?


It is.

Many gave up well considered plans in the chaos. Many sold low and many bought low. There were winners and losers.

Next time the sets of folks could be the reversed. Like in the early thirties.

So there is one happy example and one sad example. Can you accurately predict the outcome of the next example?

Again, might be good, might be bad. Do not overestimate your ability to predict which it will be. Just a caution. By all means feel free to do as you wish.

Since it might work out well next time I'll buy extra in the crash. Say with half my money I'll go from 60/40 to 80/20

Of course since it might work out badly, with the other half I'll go 40/60.

:)
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 12:52 pm

Rodc wrote:
If there is a rebalancing bonus, then it seems that the bonus would be even bigger if you "over-rebalance", no?


There is no consistent rebalancing bonus. Many threads on the subject so won't explain here, other than to point out that over time most rebalancing is from high growth to low growth so that high growth does not become too high an allocation; ie from stocks to bonds which lowers returns (and risk which is the true point of rebalancing).


Could you please provide some links to these threads? I didn't find anything on this particular topic.

Thanks!
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 12:55 pm

Rodc wrote:
Why isn't that timing?


It is.

Many gave up well considered plans in the chaos. Many sold low and many bought low. There were winners and losers.

Next time the sets of folks could be the reversed. Like in the early thirties.

So there is one happy example and one sad example. Can you accurately predict the outcome of the next example?

Again, might be good, might be bad. Do not overestimate your ability to predict which it will be. Just a caution. By all means feel free to do as you wish.

Since it might work out well next time I'll buy extra in the crash. Say with half my money I'll go from 60/40 to 80/20

Of course since it might work out badly, with the other half I'll go 40/60.

:)


That would be a funny tactic :).

But I feel I am writing on the opposite. When you refer to "many gave up well considered plans in the chaos", I think you are referring to the fact that a lot of people sold because "this time is different". They thought the end of the equity world was near, and said goodbye.

I am proposing to overweight/tilt your balance, because this time isn't different, stocks will recover someday. Isn't that the reason why we invest? We have some level of confidence, and despite the crash, that confidence hasn't changed. We still believe in equities, and if we have a longer horizon, it makes sense to invest in them since we are down and therefore expect a higher than average return in the future?
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Chan_va » Tue Jul 16, 2013 1:01 pm

Basically, your theory is - buy low, sell high. Sound in principle - IF you knew for certain that low was low, and that high was high.

It is easy to backtest your theory. Run the numbers are let us know if you come out ahead of the index.

Just know that you can create 100 models that will beat the index based on historical data. Fortunes have been lost trying to turn those models into actual profits. But who knows - maybe you have hit on something that the others haven't - the only way you will know is if you implement it.
User avatar
Chan_va
 
Posts: 409
Joined: Wed Dec 05, 2012 8:15 pm

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 1:08 pm

Chan_va wrote:Basically, your theory is - buy low, sell high. Sound in principle - IF you knew for certain that low was low, and that high was high.


Returning your remark: why do you re-balance each year? Isn't that to for the same reason: you buy low, with the confidence that it will go up someday.

So why is tilting your balance after a severe downturn wrong then?
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Clearly_Irrational » Tue Jul 16, 2013 1:14 pm

YDNAL wrote:How do you know when the market stops "collapsing" ?


Obviously there is no way to know for sure, however based on previous events you'd have reasonably good odds betting that most crashes will max out at a 50% loss.
User avatar
Clearly_Irrational
 
Posts: 2463
Joined: Thu Oct 13, 2011 4:43 pm

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 1:23 pm

AppelSienSapFrietjes wrote:
Chan_va wrote:Basically, your theory is - buy low, sell high. Sound in principle - IF you knew for certain that low was low, and that high was high.


Returning your remark: why do you re-balance each year? Isn't that to for the same reason: you buy low, with the confidence that it will go up someday.

So why is tilting your balance after a severe downturn wrong then?


One should not rebalance for increased returns. Mostly you rebalance to sell stocks to buy bunds which is very unlikely to increase returns (since stock tend to grow more than bonds, most of the time when you rebalance it is stocks that are too high). Rather one rebalances to return a portfolio to a desired level of risk. Lower risk if stocks are too high (which lowers expected returns, at least in general, and is most common) and increase risk if stocks are too low (which expectantly increases returns over not rebalancing).

You seem to be missing that there are two considerations: returns and risk. You focus on returns and ignore risk. At least seemingly.
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby bertilak » Tue Jul 16, 2013 1:34 pm

AppelSienSapFrietjes wrote:A few threads above and years ago, Rick Ferri was shouting in 2008 that young people should go all-in for equities ("Early Savers (20s and 30s) - buy equities index funds like crazy with what you can and do not look at your account balance for 10 years."). W. Buffet was making the same remarks in 2008.

Why isn't that timing?

Sure it's timing, but it is not market timing -- it is life-cycle timing. Younger investors have more time and resources (i.e. income) to recover from bad times so are justified in taking more risk. Also, early on, there is less money being put at risk. (I guess that's saying almost the same thing.)
No-one really listens to anyone else, and if you try it sometime you will see why. | -- Mignon McLaughlin
User avatar
bertilak
 
Posts: 3176
Joined: Tue Aug 02, 2011 6:23 pm
Location: East of the Pecos, West of the Mississippi

Re: Dynamic AA - does it make sense?

Postby Kevin M » Tue Jul 16, 2013 1:40 pm

OP, as mentioned in at least one reply, what you are proposing has been referred to as "overbalancing". One of our BH members, tfb, who publishes a very interesting blog, did this during the 2008/2009 crisis, and published what he planned to do and what he did as he went. I think you will find this very interesting and informative. Do this Google search to find the blog posts: overbalancing site:thefinancebuff.com.

I personally did not have the courage (some might call it foolishness) to do this. I didn't even fully rebalance back to my AA, but I did at least continue to add a fair amount to equities. I guess I was more in the camp of "what if we really are going into a long depression?"

Being retired, I didn't think I could afford to take the risk of continuing to dump too much money into equities as they plunged. I think if you are young, with a portfolio that is very small compared to your human capital, it might be an OK thing to do. But as others have pointed out, job/employment stability is a factor to consider.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4086
Joined: Mon Jun 29, 2009 4:24 pm

Re: Dynamic AA - does it make sense?

Postby entyrii » Tue Jul 16, 2013 1:43 pm

AppelSienSapFrietjes wrote: To begin with, I don't think this is market timing, since you are over weighting assets which have done bad in the past. ... Does this make sense? What is your opinion?


I'm a boglehead, so by definition I don't believe in market timing. Adjusting your AA based on PAST performance with the intention of maximizing FUTURE returns and beat the market based on any criteria is absolutely, unabashedly, market timing.

So to answer your question as a boglehead: No, this does not make sense. My opinion is quite negative. At best I doubt you would outperform buy-and-hold in the long term. At worst, your timing criteria/mechanism could misfire so to speak and get you to shift allocations at exactly the wrong moment. Not worth the time and effort to implement.
"The more they overthink the plumbing, the easier it is to stop up the drain." | -Montgomery Scott
User avatar
entyrii
 
Posts: 46
Joined: Tue Sep 25, 2012 6:38 pm

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 1:45 pm

Thanks, I will definitely trying to find that blog!

I agree that age (that is, years to retirement) and size of portfolio with respect to future savings is an important factor.

(For your interest, I am 30, so maybe I am looking this from another perspective than the default user on this forum. However, during my last 8 working years, I managed to save around 80% of my main income. So my capital is higher than an average 30y old person.)
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby FNK » Tue Jul 16, 2013 1:47 pm

Look up threads on PE10 - it does what you say in a more disciplined fashion. If you're willing to handle the complexity and added risk, of course.
User avatar
FNK
 
Posts: 1353
Joined: Tue May 17, 2011 8:01 pm

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 1:51 pm

entyrii wrote:
AppelSienSapFrietjes wrote:So to answer your question as a boglehead: No, this does not make sense. My opinion is quite negative. At best I doubt you would outperform buy-and-hold in the long term. At worst, your timing criteria/mechanism could misfire so to speak and get you to shift allocations at exactly the wrong moment. Not worth the time and effort to implement.


Thanks for your reply. I would like to address your comment mathematically (and also hypothetically, I am not proposing this).

Say your AA is 70/30. The market crashes. It is down -30% so you decide to move to 100/0. You wait, wait, wait. The market goes up and down. Suddenly, the market starts recovering. First, the market enters the same point on which you moved from 70/30 to 100/0, but now a few years later. You keep your 100/0 AA. You still wait, wait, wait. The market goes up and down. Then, the market now rises. At a certain point, you decide to move again from 100/0 to 70/30.

Where did you loose when comparing to buy-and-hold and re-balancing as usual?

Thanks!
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 1:56 pm

For your interest, here is the blog mentioned above: http://thefinancebuff.com/embracing-bear-market.html.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 1:56 pm

FNK wrote:Look up threads on PE10 - it does what you say in a more disciplined fashion. If you're willing to handle the complexity and added risk, of course.


Will do! I will mention it here once I have found it. Thanks for pointer.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby nisiprius » Tue Jul 16, 2013 1:59 pm

The biggest argument against dynamic asset allocation--usually called "tactical asset allocation"--is tactical asset allocation mutual funds.

What, you say? Right. You don't hear much about them, do you? Well, they were all the rage fifteen or twenty years ago. They were funds that did exactly what you suggest. They declared a "neutral" allocation for the fund, but adjusted the allocation up and down within a range around that "neutral" point. For example, Fidelity's Asset Manager 60% says that among its strategies are "Maintaining a neutral mix over time of 60% of assets in stocks, 35% of assets in bonds, and 5% of assets in short-term and money market instruments; adjusting allocation among asset classes gradually within the following ranges: stock class (40%-90%), bond class (10%-60%), and short-term/money market class (0%-50%)." In Vanguard's case, they claimed to be using a "quantitative model" to make the adjustments. Other fund companies might say they are doing it based on "the fund advisors' projections of returns within the asset classes."

Why don't you hear much about them? Well, let's compare Fidelity's Asset Manager, with it's "neutral mix" of 60% stocks, with Vanguard Balanced Index, which simply holds 60% in stocks and the heck with it, no fiddling.

Image

The brilliant Fidelity managers evidently managed to adjust the asset allocation at just the wrong time so as to achieve more risk, in the form of a bigger tumble in 2008-2009, yet less return, overall.

The asset allocation funds pretty much failed and went out of fashion. Vanguard used to include theirs as one of the components in their LifeStrategy funds, but they have now frozen the allocation in their "asset allocation fund" and are in the process of killing it off.

If the professional fund managers haven't succeeded with tactical asset allocation, it must not be all that easy to do.
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
nisiprius
Advisory Board
 
Posts: 25236
Joined: Thu Jul 26, 2007 10:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Dynamic AA - does it make sense?

Postby YDNAL » Tue Jul 16, 2013 2:02 pm

Clearly_Irrational wrote:
YDNAL wrote:How do you know when the market stops "collapsing" ?

Obviously there is no way to know for sure, however based on previous events you'd have reasonably good odds betting that most crashes will max out at a 50% loss.

Think again ! There were multiple upswings, followed by downswings, down to near 90% drop in 3 years.
Image
Landy | Be yourself, everyone else is already taken -- Oscar Wilde
YDNAL
 
Posts: 13404
Joined: Tue Apr 10, 2007 5:04 pm
Location: Biscayne Bay

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 2:07 pm

nisiprius wrote:The biggest argument against dynamic asset allocation--usually called "tactical asset allocation"--is tactical asset allocation mutual funds.

What, you say? Right. You don't hear much about them, do you? Well, they were all the rage fifteen or twenty years ago. They were funds that did exactly what you suggest. They declared a "neutral" allocation for the fund, but adjusted the allocation up and down within a range around that "neutral" point. For example, Fidelity's Asset Manager 60% says that among its strategies are "Maintaining a neutral mix over time of 60% of assets in stocks, 35% of assets in bonds, and 5% of assets in short-term and money market instruments; adjusting allocation among asset classes gradually within the following ranges: stock class (40%-90%), bond class (10%-60%), and short-term/money market class (0%-50%)." In Vanguard's case, they claimed to be using a "quantitative model" to make the adjustments. Other fund companies might say they are doing it based on "the fund advisors' projections of returns within the asset classes."

Why don't you hear much about them? Well, let's compare Fidelity's Asset Manager, with it's "neutral mix" of 60% stocks, with Vanguard Balanced Index, which simply holds 60% in stocks and the heck with it, no fiddling.

Image

The brilliant Fidelity managers evidently managed to adjust the asset allocation at just the wrong time so as to achieve more risk, in the form of a bigger tumble in 2008-2009, yet less return, overall.

The asset allocation funds pretty much failed and went out of fashion. Vanguard used to include theirs as one of the components in their LifeStrategy funds, but they have now frozen the allocation in their "asset allocation fund" and are in the process of killing it off.

If the professional fund managers haven't succeeded with tactical asset allocation, it must not be all that easy to do.


Very interesting. Thanks!

But the Fidelity's Asset Manager is a more extreme example of what I am suggesting. They appear to have ranges for equities (40 to 90%) and bonds (10 to 60%). So they are timing equities/bonds in favor of bonds/equities, whenever they like. That is in my knowledge the purest example of timing. Most active funds with this kind of freedom (no fixed asset/bonds allocation) do something like this, if I am not mistaken.

I am only over-weighing equities on their downturn and returning back to the default AA when equities are stabilized, in the knowledge (based on previous events)/confidence that equities must stabilize some day.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 2:09 pm

YDNAL wrote:
Clearly_Irrational wrote:
YDNAL wrote:How do you know when the market stops "collapsing" ?

Obviously there is no way to know for sure, however based on previous events you'd have reasonably good odds betting that most crashes will max out at a 50% loss.

Think again ! There were multiple upswings, followed by downswings, down to near 90% drop in 3 years.
Image


Very true. But in my opinion, that is no real counter argument. If anything, the recovery duration has been shortened. That is, if you only rebalanced according to your original AA, you had to wait X years for recovery. If you over-balanced somewhere along the way, you had to wait X-Y years for recovery.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Clearly_Irrational » Tue Jul 16, 2013 2:23 pm

YDNAL wrote:
Clearly_Irrational wrote:
YDNAL wrote:How do you know when the market stops "collapsing" ?

Obviously there is no way to know for sure, however based on previous events you'd have reasonably good odds betting that most crashes will max out at a 50% loss.

Think again ! There were multiple upswings, followed by downswings, down to near 90% drop in 3 years.


Yes, I was well aware of the great crash and the fact that it was over 50%, that doesn't invalidate my statement.
User avatar
Clearly_Irrational
 
Posts: 2463
Joined: Thu Oct 13, 2011 4:43 pm

Re: Dynamic AA - does it make sense?

Postby rkhusky » Tue Jul 16, 2013 2:26 pm

I do something similar to what you are suggesting. I have a spreadsheet that computes my AA glide path from a few years ago up to retirement. I adjust my AA with a linear function such that if the market drops by 50%, I increase the first year's stock allocation by 5% (e.g. 70% -> 75% stock), with a maximum adjustment of 5%. This means that as I get closer to retirement, the largest possible adjustment to my baseline AA will get smaller. This is all done automatically within my rebalancing calculations. And I realize I am making such a small change that it likely won't amount to much, but it scratches the itch that I should be buying more during a drop (apart from normal rebalancing).
rkhusky
 
Posts: 1568
Joined: Thu Aug 18, 2011 9:09 pm

Re: Dynamic AA - does it make sense?

Postby Chan_va » Tue Jul 16, 2013 2:31 pm

OP,

State your model in exact terms. What condition would drive an overbalance? When would you return to your default allocation? If you use a % drop as a criteria, what is the starting date from when you would calculate the % drop? If you are using PE10, what exact numbers would you use to move in and out of your allocation?

Your posts are full of hypothesis of what would happen if you bought low and sold high. Boil them down to hard models and we can backtest.
User avatar
Chan_va
 
Posts: 409
Joined: Wed Dec 05, 2012 8:15 pm

Re: Dynamic AA - does it make sense?

Postby FNK » Tue Jul 16, 2013 2:34 pm

AppelSienSapFrietjes wrote:Say your AA is 70/30. The market crashes. It is down -30% so you decide to move to 100/0. You wait, wait, wait.

Simpler! Set your AA to 100/0 and stay there. The market will go up and down, but so far the expected return from stocks is higher than from bonds, so you'll come out ahead.

Just make sure not to retire in 2008.
User avatar
FNK
 
Posts: 1353
Joined: Tue May 17, 2011 8:01 pm

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 2:38 pm

Chan_va wrote:OP,

State your model in exact terms. What condition would drive an overbalance? When would you return to your default allocation? If you use a % drop as a criteria, what is the starting date from when you would calculate the % drop? If you are using PE10, what exact numbers would you use to move in and out of your allocation?

Your posts are full of hypothesis of what would happen if you bought low and sold high. Boil them down to hard models and we can backtest.


I have no applications to back test these kind of things. If you are able to backtest this over a long period (70s to now? 20s to now?), that would be very cool!

Let's assume the following to start with: 70/30 by default. Whenever equities are down 25%, you move to 100/0. Once they are again 33.33% above that point (which means they have stabilized), you revert to 70/30.

I have no idea if this would be my model. Probably not, since 70/30 -> 100/0 is heavy. But it is interesting to start here, in my opinion. Suggestions always welcome!

Interesting!
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 2:45 pm

FNK wrote:
AppelSienSapFrietjes wrote:Say your AA is 70/30. The market crashes. It is down -30% so you decide to move to 100/0. You wait, wait, wait.

Simpler! Set your AA to 100/0 and stay there. The market will go up and down, but so far the expected return from stocks is higher than from bonds, so you'll come out ahead.


I am not sure. It depends a lot on your time window. According to Simba's spreadsheet (1972-2012):

1) TSM 100% - CAGR 9.68%
2) TSM 70%/TBM 30% - CAGR 9.45%

If you managed to apply this model once in 2008 on the second portfolio and gain another 10% (which you will if you go to 100/0 when the market drops 25% and go again to 70/30 when the market recovers from this drop), your CAGR would have been:

((1.0945^41)*1.10)^(1/41) = 9.70%.

edit: This is not really correct. This assumes that bonds were flat during this recovery. So, 10% is the upper limit. We will have to backtest it with real data.


The time window is important. If there are no opportunities to apply this model (for example during 82 to 99), it is better to be 100% in TSM. But that's not really the discussion, I believe.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Kevin M » Tue Jul 16, 2013 2:55 pm

OP, you posted the link to the blog post where tfb started the process. It's also interesting to read some of the other posts, including the one where he concluded the process after recovery: Take Money Off the Table After a Good Run. It worked for him, but he admitted he was lucky.

Kevin
||.......|| Suggested format for Asking Portfolio Questions (edit original post)
User avatar
Kevin M
 
Posts: 4086
Joined: Mon Jun 29, 2009 4:24 pm

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 3:07 pm

http://home.comcast.net/~rodec/finance/ ... mingV2.pdf

Conclusions:
The initial attempts to use q as the basis for a “valuation matters” algorithm to improve portfolio performance failed. However, when the portfolio switching is stabilized by time smoothing q, the “valuation matters” approach did provide benefits, beating buy and hold frequently (72 wins to 17 losses) and by a solid margin (approximately 14% or 25% in total dollars over an investment period of 20 years).

Interestingly, and perhaps unfortunately, the vast majority of the out-performance came in approximately the first 50 of the 89 rolling 20-year periods available. Very little if any benefit has been provided for nearly 40 years. The q-timing approach as examined here, with or without smoothing would have failed to provide protection in the recent downturn. While one can certainly imagine other timing algorithms based on q that would have pulled an investor out of stocks prior to this downturn, it would remain to show an over-all benefit, and perhaps more importantly, one has to fear that classic downfall of so many investing methodologies, data mining.

So, the bottom line as I see it is that using q to adjust your portfolio may provide some benefit going forward, or given the results over nearly the last 40 years, it may not. But, when the market turns against investors, I see little salvation in q. In retrospect this is not surprising. Market returns are partly rational, based on profitability of the underlying businesses. Market returns are also partly irrational, based on investor sentiment, for example investor optimism driving up price to earnings ratios or investor pessimism driving down price to earnings ratios. Valuation measures might be able to measure the rational part, but it is not surprising when they fail to get the irrational part correct. Valuation may matter, but timing the irrational part matters too, and it matters a lot.


The above little paper might be of interest. It looks at market timing based on Tobin's Q, supposed by some to be a better measure of over/under valuation of the stock market than P/E10. After I wrote it someone pointed out that in addition to no benefit for the 40 years ending in 2008 when I did the study (so did not work in the tech crash) it only worked on the data used to develop Q, and then immediately failed to provide a benefit.

A couple of things to remember. Anyone with an eight grade education and a spread sheet can build a timing scheme that back tests well, no real understanding of the economy, stock and bond markets, or investing is needed. If you make 100 such schemes, somewhere around 5 of them will have performance above the null hypothesis that is "statistically significant" (at the 95th percentile likelihood) even if they are just random noise! So just because someone did a lot of work to tune a timing scheme on past data is not comforting if you want to use it going forward. Which is why history is littered with schemes that looked good until someone tried them real-time.

My first attempt did not work, it used Tobin's Q directly. I tried again and with the right amount of temporal smoothing it worked a little, on the dependent data. And failed on the dependent data.

This in no way means it can't be done. I think it does mean it is a lot harder than 99.44% of the population believes. Simple hand waving about no brainers and such carry about zero weight.

Of course, perhaps someone will repeat the work including the data from the time of my work. Maybe after failing for 40 years it worked the last 4.

At the very least if you think this is easy, get a long history of returns (see Ibbotson, Shiller, or Fama/French), tune a scheme on say the first half, and then test on the second. Repeat in the reverse. One little test that is often illuminating: tune the switching parameters on the first half and the second half. If you get rather different parameters you are in trouble. If you get very similar parameters you have some hope.

Best of luck.
Last edited by Rodc on Tue Jul 16, 2013 3:09 pm, edited 1 time in total.
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 3:09 pm

Further thinking: it depends on the pace of the recovery of the equities compared to the growth of the bounds during that recovery.

Assume following model: You go from 70/30 to 70+X/30-X if equities loose -25%. If equities are again 33% above that point, you revert to 70/30.

If during recovery, the accumulated growth of X% bonds is greater than the accumulated growth of X% equities, you loose. Else, you win. In other words, the faster the recovery of the equities, the better.
Last edited by AppelSienSapFrietjes on Tue Jul 16, 2013 3:12 pm, edited 1 time in total.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 3:12 pm

I don't think you need to convince people that a faster recover is better than a slow one. Or slow is better than never.
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby Chan_va » Tue Jul 16, 2013 3:20 pm

Appel, here is the sequence of events that investors with some math bent go through

1. Build Model X, back test against some data and realize it does much better than the average
2. Become excited, and implement it till X fails
3. Determine that Model X was not sophisticated enough, and tune it to account for failure in #2
4. Repeat #1-#3 n times
5. Give up and spend free time gardening

There is no real way to convince anyone till they have gone through steps 1-5, so good luck.
User avatar
Chan_va
 
Posts: 409
Joined: Wed Dec 05, 2012 8:15 pm

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 3:21 pm

Rodc wrote:http://home.comcast.net/~rodec/finance/papers/Q-timingV2.pdf

Conclusions:
The initial attempts to use q as the basis for a “valuation matters” algorithm to improve portfolio performance failed. However, when the portfolio switching is stabilized by time smoothing q, the “valuation matters” approach did provide benefits, beating buy and hold frequently (72 wins to 17 losses) and by a solid margin (approximately 14% or 25% in total dollars over an investment period of 20 years).

Interestingly, and perhaps unfortunately, the vast majority of the out-performance came in approximately the first 50 of the 89 rolling 20-year periods available. Very little if any benefit has been provided for nearly 40 years. The q-timing approach as examined here, with or without smoothing would have failed to provide protection in the recent downturn. While one can certainly imagine other timing algorithms based on q that would have pulled an investor out of stocks prior to this downturn, it would remain to show an over-all benefit, and perhaps more importantly, one has to fear that classic downfall of so many investing methodologies, data mining.

So, the bottom line as I see it is that using q to adjust your portfolio may provide some benefit going forward, or given the results over nearly the last 40 years, it may not. But, when the market turns against investors, I see little salvation in q. In retrospect this is not surprising. Market returns are partly rational, based on profitability of the underlying businesses. Market returns are also partly irrational, based on investor sentiment, for example investor optimism driving up price to earnings ratios or investor pessimism driving down price to earnings ratios. Valuation measures might be able to measure the rational part, but it is not surprising when they fail to get the irrational part correct. Valuation may matter, but timing the irrational part matters too, and it matters a lot.


The above little paper might be of interest. It looks at market timing based on Tobin's Q, supposed by some to be a better measure of over/under valuation of the stock market than P/E10. After I wrote it someone pointed out that in addition to no benefit for the 40 years ending in 2008 when I did the study (so did not work in the tech crash) it only worked on the data used to develop Q, and then immediately failed to provide a benefit.

A couple of things to remember. Anyone with an eight grade education and a spread sheet can build a timing scheme that back tests well, no real understanding of the economy, stock and bond markets, or investing is needed. If you make 100 such schemes, somewhere around 5 of them will have performance above the null hypothesis that is "statistically significant" (at the 95th percentile likelihood) even if they are just random noise! So just because someone did a lot of work to tune a timing scheme on past data is not comforting if you want to use it going forward. Which is why history is littered with schemes that looked good until someone tried them real-time.

My first attempt did not work, it used Tobin's Q directly. I tried again and with the right amount of temporal smoothing it worked a little, on the dependent data. And failed on the dependent data.

This in no way means it can't be done. I think it does mean it is a lot harder than 99.44% of the population believes. Simple hand waving about no brainers and such carry about zero weight.

Of course, perhaps someone will repeat the work including the data from the time of my work. Maybe after failing for 40 years it worked the last 4.

At the very least if you think this is easy, get a long history of returns (see Ibbotson, Shiller, or Fama/French), tune a scheme on say the first half, and then test on the second. Repeat in the reverse. One little test that is often illuminating: tune the switching parameters on the first half and the second half. If you get rather different parameters you are in trouble. If you get very similar parameters you have some hope.

Best of luck.


Thanks. I will read it through and let you know. Very interesting.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby livesoft » Tue Jul 16, 2013 3:33 pm

I love a certain kind of market timing. Just search the forum for "livesoft RBD" and enjoy reading all the discussions.

As for rebalancing, all the published studies don't rebalance in an optimal way, so of course they don't show that rebalancing is a big win. In contrast, I try to rebalance only at the market highs and the market lows each year. What if you rebalanced optimally? Would rebalancing then show a big win?

And finally, if someone has a great method of rebalancing optimally, are they going to tell everyone about it? Or are they going to obfuscate what they do?
It's all about short-term opportunistic rebalancing due to a short-term change in one's asset allocation, uh, I mean opportunistic rebalancing, uh I mean rebalancing, uh I mean market timing.
livesoft
 
Posts: 33243
Joined: Thu Mar 01, 2007 9:00 pm

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 3:47 pm

livesoft wrote:And finally, if someone has a great method of rebalancing optimally, are they going to tell everyone about it? Or are they going to obfuscate what they do?


Neither. Open a fund and charge 120% of the benefit for management fees. :)
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby Chris M » Tue Jul 16, 2013 4:15 pm

AppelSienSapFrietjes wrote:
Rodc wrote:
If there is a rebalancing bonus, then it seems that the bonus would be even bigger if you "over-rebalance", no?


There is no consistent rebalancing bonus. Many threads on the subject so won't explain here, other than to point out that over time most rebalancing is from high growth to low growth so that high growth does not become too high an allocation; ie from stocks to bonds which lowers returns (and risk which is the true point of rebalancing).


Could you please provide some links to these threads? I didn't find anything on this particular topic.

Thanks!


William Bernstein has an article on his website http://www.efficientfrontier.com/ef/996/rebal.htm in which he derives an equation for the rebalancing bonus, and tests it over the period 1988 to 1994 using a variety of asset pairs. He gets an R-squared of 0.98 using his equation. Bernstein says "when looking at the 1970-94 period, rebalancing various asset pairs almost always provides returns superior to nonrebalanced porfolios." He also says that for a 50/50 mix of stocks and long-term corporate bonds, over the period 1926-1994, a non-rebalanced portfolio would out-return a rebalanced portfolio, because the portfolio drifts away from 50/50 and winds up 90 percent stock. But he points out that the higher returns of the non-rebalanced portfolio is a result of this portfolio's increasing level of risk.

I agree that the primary purpose of rebalancing is to reduce risk, but the resulting risk reduction sometimes--not always--comes with a return bonus. To the extent that rebalancing successfully reduces portfolio volatility it will enhance compound returns. Sometimes this effect will be more than offset by the tendency of rebalancing to result in a net transfer out of equities and into bonds, but sometimes the improvement in compound returns will dominate. You would expect the former; the latter is surprising. Even the possibility of an inconsistent rebalancing bonus is remarkable, since it is coming with reduced volatility. You shouldn't expect it, but you might get it--which is all the more motivation to rebalance when you really don't want to, like in the middle of a major selloff.
Chris M
 
Posts: 88
Joined: Tue May 28, 2013 7:25 pm

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 4:19 pm

Chris M wrote:
AppelSienSapFrietjes wrote:
Rodc wrote:
If there is a rebalancing bonus, then it seems that the bonus would be even bigger if you "over-rebalance", no?


There is no consistent rebalancing bonus. Many threads on the subject so won't explain here, other than to point out that over time most rebalancing is from high growth to low growth so that high growth does not become too high an allocation; ie from stocks to bonds which lowers returns (and risk which is the true point of rebalancing).


Could you please provide some links to these threads? I didn't find anything on this particular topic.

Thanks!


William Bernstein has an article on his website http://www.efficientfrontier.com/ef/996/rebal.htm in which he derives an equation for the rebalancing bonus, and tests it over the period 1988 to 1994 using a variety of asset pairs. He gets an R-squared of 0.98 using his equation. Bernstein says "when looking at the 1970-94 period, rebalancing various asset pairs almost always provides returns superior to nonrebalanced porfolios." He also says that for a 50/50 mix of stocks and long-term corporate bonds, over the period 1926-1994, a non-rebalanced portfolio would out-return a rebalanced portfolio, because the portfolio drifts away from 50/50 and winds up 90 percent stock. But he points out that the higher returns of the non-rebalanced portfolio is a result of this portfolio's increasing level of risk.

I agree that the primary purpose of rebalancing is to reduce risk, but the resulting risk reduction sometimes--not always--comes with a return bonus. To the extent that rebalancing successfully reduces portfolio volatility it will enhance compound returns. Sometimes this effect will be more than offset by the tendency of rebalancing to result in a net transfer out of equities and into bonds, but sometimes the improvement in compound returns will dominate. You would expect the former; the latter is surprising. Even the possibility of an inconsistent rebalancing bonus is remarkable, since it is coming with reduced volatility. You shouldn't expect it, but you might get it--which is all the more motivation to rebalance when you really don't want to, like in the middle of a major selloff.


Is that the article he later retracted?

ADDED: a quick search on the web turned up this 2012 quote from Bernstein posting here on Bogleheads:

I tend to think these days in terms of risky assets and riskless ones as the only two asset classes. As others above have pointed out, you generally don't earn excess return rebalancing between stocks and bonds simply because stocks have so much higher expected and realized returns.


viewtopic.php?f=10&t=94071#p1354353
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby Chris M » Tue Jul 16, 2013 4:24 pm

Rodc wrote:
Chris M wrote:
AppelSienSapFrietjes wrote:
Rodc wrote:
If there is a rebalancing bonus, then it seems that the bonus would be even bigger if you "over-rebalance", no?


There is no consistent rebalancing bonus. Many threads on the subject so won't explain here, other than to point out that over time most rebalancing is from high growth to low growth so that high growth does not become too high an allocation; ie from stocks to bonds which lowers returns (and risk which is the true point of rebalancing).


Could you please provide some links to these threads? I didn't find anything on this particular topic.

Thanks!


William Bernstein has an article on his website http://www.efficientfrontier.com/ef/996/rebal.htm in which he derives an equation for the rebalancing bonus, and tests it over the period 1988 to 1994 using a variety of asset pairs. He gets an R-squared of 0.98 using his equation. Bernstein says "when looking at the 1970-94 period, rebalancing various asset pairs almost always provides returns superior to nonrebalanced porfolios." He also says that for a 50/50 mix of stocks and long-term corporate bonds, over the period 1926-1994, a non-rebalanced portfolio would out-return a rebalanced portfolio, because the portfolio drifts away from 50/50 and winds up 90 percent stock. But he points out that the higher returns of the non-rebalanced portfolio is a result of this portfolio's increasing level of risk.

I agree that the primary purpose of rebalancing is to reduce risk, but the resulting risk reduction sometimes--not always--comes with a return bonus. To the extent that rebalancing successfully reduces portfolio volatility it will enhance compound returns. Sometimes this effect will be more than offset by the tendency of rebalancing to result in a net transfer out of equities and into bonds, but sometimes the improvement in compound returns will dominate. You would expect the former; the latter is surprising. Even the possibility of an inconsistent rebalancing bonus is remarkable, since it is coming with reduced volatility. You shouldn't expect it, but you might get it--which is all the more motivation to rebalance when you really don't want to, like in the middle of a major selloff.


Is that the article he later retracted?


I don't know, is it?--it's still on his website.
Chris M
 
Posts: 88
Joined: Tue May 28, 2013 7:25 pm

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 4:26 pm

Chris, see the edit to my post above. I was not quick enough and posted while you were posting.
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 4:34 pm

I have written a quick and dirty Java application. I have stolen data from Simba's spreadsheet. It ranges from 1972 to 2012.

The default AA is 70/30. If the market tanks 25%, I switch to 100/0. If the market recovers (which means it has to rise 33% beginning from previous point), I restore AA to 70/30.

This is the output:

1: 4169.3385594629235
Tiling starts at year 1974!
Tiling stops at year 1975!
Tiling starts at year 2002!
Tiling stops at year 2004!
Tiling starts at year 2008!
Tiling stops at year 2010!
2: 5285.499660486326

This means that 100$ (70+30) went to 4169$ in the default rebalancing setup. In the model above, it went to 5285$.

This does not take inflation into account.

I want to stress out that this outcome is preliminary. I have to verify my code. I am out for a run, so I can clear my head and think about the correctness. I am happy to distribute the source code.

Let me know if you want to see the output of other models. I can easily change:

1) default AA (70/30)
2) tilt AA (100/0)
3) threshold to change from default AA to tilt AA (-25%)
3) threshold to change from tilt AA to default AA (+33%)
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Re: Dynamic AA - does it make sense?

Postby Chris M » Tue Jul 16, 2013 4:39 pm

Rodc wrote:Chris, see the edit to my post above. I was not quick enough and posted while you were posting.


Thanks for the link Rodc. He does make the same point in the article too, that rebalancing is more likely to yield a bonus in a global equity portfolio (where long-run returns among asset classes are more or less similar) than across stocks and bonds. He says that if the difference in long-term returns between the assets being rebalanced exceeds 5%, then non-rebalanced portfolios will win out.
Chris M
 
Posts: 88
Joined: Tue May 28, 2013 7:25 pm

Re: Dynamic AA - does it make sense?

Postby Rodc » Tue Jul 16, 2013 5:03 pm

AppelSienSapFrietjes wrote:I have written a quick and dirty Java application. I have stolen data from Simba's spreadsheet. It ranges from 1972 to 2012.

The default AA is 70/30. If the market tanks 25%, I switch to 100/0. If the market recovers (which means it has to rise 33% beginning from previous point), I restore AA to 70/30.

This is the output:

1: 4169.3385594629235
Tiling starts at year 1974!
Tiling stops at year 1975!
Tiling starts at year 2002!
Tiling stops at year 2004!
Tiling starts at year 2008!
Tiling stops at year 2010!
2: 5285.499660486326

This means that 100$ (70+30) went to 4169$ in the default rebalancing setup. In the model above, it went to 5285$.

This does not take inflation into account.

I want to stress out that this outcome is preliminary. I have to verify my code. I am out for a run, so I can clear my head and think about the correctness. I am happy to distribute the source code.

Let me know if you want to see the output of other models. I can easily change:

1) default AA (70/30)
2) tilt AA (100/0)
3) threshold to change from default AA to tilt AA (-25%)
3) threshold to change from tilt AA to default AA (+33%)


That is not really kosher. One has an average of 70% stocks and the other has a higher percentage in stocks. Thus one has more risk, so it should have more return. Does it also have a better risk adjusted return?

The easiest way to do this is run your market timing portfolio. Figure out the average stock allocation. Then make the fixed allocation portfolio match. This then roughly matches up risk, so you can compare returns. (not a perfect solution to the problem but much better).

If you look under the hood I bet the results are almost entirely driven by the last tilt period. I also note somehow it misses the great run up of the 1990s, which is not a good thing. Do you want something that works because of one and only one fortunate bit of timing on the greatest crash and recovery in 80 years, while it missed its greatest opportunity the great Tech run up?

It is also worth the caveat that is amounts to only one data point over a period of about one investing lifetime. While interesting it is of limited value in proving something works.

If you are so inclined I would suggest getting a much longer set of data and try the experiments I suggested earlier.
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.
Rodc
 
Posts: 10111
Joined: Tue Jun 26, 2007 10:46 am

Re: Dynamic AA - does it make sense?

Postby AppelSienSapFrietjes » Tue Jul 16, 2013 5:11 pm

Rodc wrote:
AppelSienSapFrietjes wrote:I have written a quick and dirty Java application. I have stolen data from Simba's spreadsheet. It ranges from 1972 to 2012.

The default AA is 70/30. If the market tanks 25%, I switch to 100/0. If the market recovers (which means it has to rise 33% beginning from previous point), I restore AA to 70/30.

This is the output:

1: 4169.3385594629235
Tiling starts at year 1974!
Tiling stops at year 1975!
Tiling starts at year 2002!
Tiling stops at year 2004!
Tiling starts at year 2008!
Tiling stops at year 2010!
2: 5285.499660486326

This means that 100$ (70+30) went to 4169$ in the default rebalancing setup. In the model above, it went to 5285$.

This does not take inflation into account.

I want to stress out that this outcome is preliminary. I have to verify my code. I am out for a run, so I can clear my head and think about the correctness. I am happy to distribute the source code.

Let me know if you want to see the output of other models. I can easily change:

1) default AA (70/30)
2) tilt AA (100/0)
3) threshold to change from default AA to tilt AA (-25%)
3) threshold to change from tilt AA to default AA (+33%)


That is not really kosher. One has an average of 70% stocks and the other has a higher percentage in stocks. Thus one has more risk, so it should have more return. Does it also have a better risk adjusted return?

The easiest way to do this is run your market timing portfolio. Figure out the average stock allocation. Then make the fixed allocation portfolio match. This then roughly matches up risk, so you can compare returns. (not a perfect solution to the problem but much better).

It is also worth the caveat that is amounts to only one data point over a period of about one investing lifetime. While interesting it is of limited value in proving something works.

If you are so inclined I would suggest getting a much longer set of data and try the experiments I suggested earlier.


Easy :) I need more time to plug everything in. This was only a quick test to determine how easy it is to write a simulation in Java. I finished in about 1.5 hours, so that's OK.

- I have to include risk calculation. I have to think about how this should be done.
- I should include more data. Though, I have no idea where to get this in downloadable/parsable format.
- This test goes from 1972 to 2012. I should include any configurable time window.
AppelSienSapFrietjes
 
Posts: 93
Joined: Thu Jun 20, 2013 6:50 am

Next

Return to Investing - Theory, News & General

Who is online

Users browsing this forum: BigJohn, BL, Clearly_Irrational, Day9, Google [Bot], madbrain, Medviersi, mickeyd, noktem, riderdie4life, stemikger, Tamales, Taylor Larimore, z3r0c00l and 78 guests