rebalancing PP since 1972

[modeljc]

Has anyone tested back to 1972 to see how many times you would need to rebalance using 35%/15% and 30%/20%?  I assume that 35%/15% had the highest returns.

[moda0306]

I think Clive had something on here but he deletes his old posts.

You are correct... 35/15 has the best returns.

[stone]

moda, do I remember rightly that the difference between 35/15 and 30/20 and annual or monthly or daily rebalancing is actually astonishingly small.

[moda0306]

stone,

I believe doing 35/15 vs annual would deliver about .5% more... I think that's what I remember hearing from someone who calculated it.

I don't know if that counts as "astonishingly small," but it's what my memory is feeding me right now.

[clacy]

Optimal rebalancing will be dependent on the time period and what the markets are doing.  With wider rebalance bands, you can catch bigger trends.  If you went into a time period that was choppier, 30/20 bands will be better.  Of course, hindsight will tell you what was the best, but there's no way to know in advance that I know of.

[bswift]

I had a similar question, how often would a retiree taking withdrawals have rebalanced using back data?  I started 1/1/76 to leave out the early gold boom, and went through 2011.  Assumptions were:

Used craigr's return data from this forum
No accounting for taxes -- too difficult
Reinvestment of divs/int so I could use component annual returns
4% initial withdrawal adjusted higher for CPI each year
Withdrawals taken from highest asset class, or from more than one if close in value, to make equal
35/15 rebalance bands

And it came out to 5.  5 times in 36 years. 

After 1979, so 1/1/80, gold was 45% of the portfolio.  That made me wonder if only checking annually might be better than rebalancing the second an asset hits 35%, or even a bit earlier as some people say here.  I may compromise and check twice a year but no more.  This was lucky, catching most of the blow-off rise.  I wonder if even PP devotees are so eager to do something, anything, that waiting until an arbitrary date, potentially past 35%, is just too hard.

1/1/84 gold was down to 14.1% and had to be rebought, at prices quite a bit lower than the 1980 sale.

1/1/92 gold was down to 13.1% and had to be bought again.  This might have been a tough buy, after a bad decade.

1/1/98 gold was down to 13.6% and stocks were exactly 35%.  The first non-gold rebalance trigger after 22 years!  Talk about a tough rebalance.  And this would have been pure torture in 1999 and early 2000.  Maybe there is an argument to be made for using a higher top band.  I mean, why does it have to be symmetrical?  If I had let stocks go to 45% in this bubble, the end result could have been spectacular sell timing again.  Too small a sample to draw any conclusions, I am sure.

Anyway it was interesting to see that the 1/1998 - 5/2000 stock bubble blow-off did not generate another rebalance.  According to my rules I was taking all of the annual withdrawals from stocks, since it had far outpaced the other three components.  That captured $50k and a bit a year.  But stock was only back to 30.9% on 1/1/2000, so no more rebalances.

Until 1/1/2009.  Stock was down to 14.8%, time to rebalance.  That was good, not perfect, timing as well.  Bonds were actually higher than gold then 31.2% to 30.6%, both would be sold to top up stocks and a bit of cash. 

This portfolio finished 2011 far from another rebalance, with gold at 30.9% and cash at 19.9% the closest to a band.  So only 5 rebalances in all that time, plus the annual mini-rebalance of withdrawing from the highest asset(s).

Incidentally, the withdrawal rate has declined from 4% on 1/1/77 to 1.86% on 1/1/12.  I doubt anyone would let it go quite that low, but I do feel like sustainable withdrawal rates have fallen closer to 2.5% as valuations have risen and interest rates have dropped.  The actual dollar amounts were $17,944 rising to $69,493.  Those sound reasonable to me, especially considering the retiree would be very old now, having been retired 36 years. 

I wonder if taking withdrawals from the highest asset, although it definitely minimized rebalances, could have hurt performance overall by muting trends?  I wouldn't do it that way in real life, simply for convenience.  I would have all dividends and interest go to cash, and sell the highest only to make up the difference in the withdrawal amount. 

[modeljc]

A great post for me bswift.  I am at the withdrawal age and was wondering about how to rebalance and also how to take a annual distribution.  I assume that the best way to take a distribution in real life is from you cash or to sell SHY down to 15% and then to rebalance.  Would that create the highest returns in your study?  I agree you would have a hard time staying with the PP during the bull market.  At age 74, it may be a little easier as you don't want a 100% stock allocation.  I do hold my noise to own TLT. 

[Alanw]

A great post for me bswift.  I am at the withdrawal age and was wondering about how to rebalance and also how to take a annual distribution.  I assume that the best way to take a distribution in real life is from you cash or to sell SHY down to 15% and then to rebalance.  Would that create the highest returns in your study?  I agree you would have a hard time staying with the PP during the bull market.  At age 74, it may be a little easier as you don't want a 100% stock allocation.  I do hold my noise to own TLT. 

modeljc
I am also at the withdrawal age.  I withdraw quarterly from a savings account which is part of my PP cash holding and replenish that as needed from my SHY holding.  My PP was set up last year, and I had to hold my nose also to purchase TLT.  Since I am also relatively new to the PP, I decided to set my rebalance bands at 30-20 the first year.  Wouldn't you know that TLT has been my biggest gainer and I already have had to rebalance, sell some TLT and purchase more stock a few months ago.  I am now contemplating moving my rebalance bands to 35-25.

[KevinW]

And it came out to 5.  5 times in 36 years. 

Yeah, the standard guidelines have you rebalancing very rarely. I think there's a tendency to worry too much about the process and  expenses of rebalancing. Except for the early part of accumulation where contributions are large relative to the total portfolio, rebalances only happen every few years.

Kind of like replacing a refrigerator. Not something that really needs to be on your mind constantly. When the time comes you'll deal with it in an afternoon and be set for another few years.

I'd be interested to see what happens if you run the same experiment but always pull from cash instead of the largest asset. My guess is that the result would be very similar.

[patrickjhall]

I had a similar question, how often would a retiree taking withdrawals have rebalanced using back data?  I started 1/1/76 to leave out the early gold boom, and went through 2011.  Assumptions were:

Used craigr's return data from this forum
No accounting for taxes -- too difficult
Reinvestment of divs/int so I could use component annual returns
4% initial withdrawal adjusted higher for CPI each year
Withdrawals taken from highest asset class, or from more than one if close in value, to make equal
35/15 rebalance bands

And it came out to 5.  5 times in 36 years.  
This portfolio finished 2011 far from another rebalance, with gold at 30.9% and cash at 19.9% the closest to a band.  So only 5 rebalances in all that time, plus the annual mini-rebalance of withdrawing from the highest asset(s).

...

Incidentally, the withdrawal rate has declined from 4% on 1/1/77 to 1.86% on 1/1/12.  I doubt anyone would let it go quite that low, but I do feel like sustainable withdrawal rates have fallen closer to 2.5% as valuations have risen and interest rates have dropped.  The actual dollar amounts were $17,944 rising to $69,493.  Those sound reasonable to me, especially considering the retiree would be very old now, having been retired 36 years.  

That is fascinating. Great post! What are the starting and ending portfolio values?

[bswift]

krigare, the portfolio started at $400,000 on 1/1/76 and ended at $3,742,581.  Past performance is no guarantee of future results.  Also, this was completely tax-free, because I couldn't possibly work out what taxes would have been on all the interest and dividends.  So the returns are seriously overstated.  I didn't think that would affect rebalance frequency very much, but it may have.

modeljc and Alanw, do you have dividends and interest reinvested into your stock fund and TLT?  Or does that get added to SHY/cash?  It seems like cash would get down to 15% pretty fast if all the withdrawals come from it, and generate much more frequent rebalances.  Are you doing that to be tax-efficient, or because cash has no yield right now, or some other reason?

Is there an HB-approved, official way of taking withdrawals in retirement? 

[modeljc]

bswift,

The dividends and interest go to cash and are not reinvested.  I don't know how to make a withdrawal and trying to figure the best way.  It seems that cash and SHY would be the most stable and one could draw it down to 15% without affecting long term performance or paying a lot of taxes.  Half of my PP is in taxable and the other half is in sheltered accounts.  I am thinking the rebalance could be done best in the sheltered accounts without creating taxes.  We also have some amount of income each year to do some rebalance in the taxable account.  In the past we take our Required Minimum Distribution RMD and do an annual rebalance in the taxable account.  So rebalance for us may not be much of a problem. 

Any suggestions on the best way to take withdrawals would be appreciated. 

[Alanw]

krigare, the portfolio started at $400,000 on 1/1/76 and ended at $3,742,581.  Past performance is no guarantee of future results.  Also, this was completely tax-free, because I couldn't possibly work out what taxes would have been on all the interest and dividends.  So the returns are seriously overstated.  I didn't think that would affect rebalance frequency very much, but it may have.

modeljc and Alanw, do you have dividends and interest reinvested into your stock fund and TLT?  Or does that get added to SHY/cash?  It seems like cash would get down to 15% pretty fast if all the withdrawals come from it, and generate much more frequent rebalances.  Are you doing that to be tax-efficient, or because cash has no yield right now, or some other reason?

Is there an HB-approved, official way of taking withdrawals in retirement? 

In my previous post, I meant to say that I am contemplating 35-15 rebalance bands rather than 30-20.  I have dividends and interest reinvested into stock and TLT accounts.  If you take a 4% withdrawal (I take 5%), it would still take 2-3 years to hit a 15% rebalance band for cash under normal circumstances.  If one of the other parts of the PP were to explode in value, it would be sooner.  I am just trying to keep my PP as simple as possible.

[stone]

Alanw, is it tax reasons that cause you to reinvest interest and dividends? Otherwise it sounds like an unfortunate way to inadvertently end up making purchases of leading assets when they are peaking. Having the dividends and interest go to cash is the standard set up isn't it?

[Alanw]

My LTT's are all in an IRA.  The stock portion is about 50-50 IRA and taxable.  It may be benefitial to have the stock dividends go to the cash account but I really hadn't given it much thought.  I'll check with my accountant this year to see if there is an advantage.  I don't know if it will make much difference when it comes to rebalancing.  Anyway, Stone, thanks for bringing it to my attention.

[escafandro]

Any suggestions on the best way to take withdrawals would be appreciated.

I subscribe to this query.
My fear is that if I'm withdrawing from cash only I'm going down artificially the asset and therefore changing the effect to buy low and sell high that gives me the 15/35 bands system.

[Alanw]

Any suggestions on the best way to take withdrawals would be appreciated.

I subscribe to this query.
My fear is that if I'm withdrawing from cash only I'm going down artificially the asset and therefore changing the effect to buy low and sell high that gives me the 15/35 bands system.

For me, the simple way is to withdraw from cash.  When you hit a 15% re-balance band with cash, you will automatically have to replenish cash with your highest returning asset to bring the portfolio back to 4X25%.