Can Portfoliocharts describe optimum rebalancing bands?

[Greg]

Tyler et Al,

viewtopic.php?f=1&t=3686&p=51581&hilit= ... lly#p51581

This has been spoken about a couple of times of what the optimum re-balancing bands are. Within your Portfoliocharts.com website (which might I add is excellent, much appreciated for making it), you have a re-balancing annually at the bottom. Not sure if you have the data, but have you looked into if the optimum re-balancing bands change based on the portfolio you go with (almost a heat map of the different re-balancing strategies and which one has the highest probability of reaching top returns)? Just curious. Thanks in advance for your thoughts (and the thoughts that come from the others on this fine forum)!

[Tyler]

Good question!

The problem is mostly about data availability. To model most rebalancing strategies other than the simple annual method (bands, momentum, etc) requires data more granular than the annual numbers I use on the site. I actually have the ability to get monthly returns now, although I haven't yet taken the time to do that. I'll have to look into it.

That said, the PP is somewhat unique in that it promotes the use of bands while only really needing to look at the numbers once a year. I'll think about it and see what it would take to mock up a more customizable chart using that method. In the meantime, you might also check out peaktotrough.com, as that offers a few different rebalancing options for the PP, at least.

[Greg]

Good question!

The problem is mostly about data availability. To model most rebalancing strategies other than the simple annual method (bands, momentum, etc) requires data more granular than the annual numbers I use on the site. I actually have the ability to get monthly returns now, although I haven't yet taken the time to do that. I'll have to look into it.

That said, the PP is somewhat unique in that it promotes the use of bands while only really needing to look at the numbers once a year. I'll think about it and see what it would take to mock up a more customizable chart using that method. In the meantime, you might also check out peaktotrough.com, as that offers a few different rebalancing options for the PP, at least.

I was playing around with peaktotrough, good stuff. Obviously there is some back-testing that can be involved that can say "for a particular investing strategy, starting at this time, 35-15 re-balancing bands provide the highest CAGR". That data is only moderately useful versus I really like your probability chunks in your portfolios that look at what's the chances of reaching this CAGR consistently, etc are.

If somehow the data could say, "70% of the time (regardless of when you'd start for a particular portfolio strategy), you'd make more money doing a 35-15 bands versus a re-balancing annually (based on a heat map of various different start dates)", that'd be some sweet data to have.

Thanks again for all you do Tyler.

[blue_ruin17]

Where's sophie?

She did some pretty elaborate backtesting with various rebalancing strategies, if I recall correctly. I believe that the 35/15 rebalancing bands produced (historically) the strongest returns.

I personally prefer contributing to the lagging asset, for a variety of reasons unrelated to historical statistics.

[sophie]

Yup, I ran some simulations comparing several different strategies for accumulation, and I remember comparing rebalancing strategies as well. 15/35 band rebalancing beat annual, and there was something on the order of 0.5% CAGR spread - quite significant. 20/30 bands did not perform as well as 15/35. Something to keep in mind with PP backtests, as they usually use either no rebalancing or annual.

Accumulation strategies were a different set of backtests, but I think I only used the 15/35 rebalancing bands for those. I ended up deciding that adding to the lagging asset was a perfectly valid choice, because it can avoid incurring fees and taxes. Also it performed only slightly worse than the winning strategy, which was to divide contributions equally among the four assets.

[Greg]

Yup, I ran some simulations comparing several different strategies for accumulation, and I remember comparing rebalancing strategies as well. 15/35 band rebalancing beat annual, and there was something on the order of 0.5% CAGR spread - quite significant. 20/30 bands did not perform as well as 15/35. Something to keep in mind with PP backtests, as they usually use either no rebalancing or annual.

Accumulation strategies were a different set of backtests, but I think I only used the 15/35 rebalancing bands for those. I ended up deciding that adding to the lagging asset was a perfectly valid choice, because it can avoid incurring fees and taxes. Also it performed only slightly worse than the winning strategy, which was to divide contributions equally among the four assets.

Interesting stuff. At least for a 25/25/25/25 portfolio, going from 25% to 35% is a 40% increase (or 25% to 15% is a 40% decrease).

I'm assuming then (hypothetically) if you'd be using something with 5 asset classes Tyler's Golden Butterfly portfolio, where you have 20/20/20/20/20 (Total Stock Market/Small-Cap Value/LT Bonds/ST Bonds/Gold) that you'd have each of those go to 40% appreciation on any one of them (20*1.4 = 28% or 20*.6 = 12%). Not sure if since both the TSM and Small-Cap would both be affected by the economic condition "prosperity" (but just in slightly different ways), if the re-balancing bands would be different. Thanks for your thoughts.

[Pet Hog]

Interesting stuff. At least for a 25/25/25/25 portfolio, going from 25% to 35% is a 40% increase (or 25% to 15% is a 40% decrease).

<pedantry>Going from 25% to 35% requires the asset to grow by about 62%. You have to remember that the whole portfolio grows along with a single asset growing. So 25 (out of 100) going to 35 (out of 110), 40% growth, means the asset has gone from 25% to about 32% of the total. It would be 35% if the asset grew from 25 (out of 100) to about 40.4 (out of 115.4), or 62% growth. On the downside, 25 (out of 100) would have to go to about 13.2 (out of 88.2) to reach 15%, or a loss of about 47%. For this reason, I don't consider the 35/15 rebalancing bands as symmetrical (the odds of rebalancing on the upside don't match those on the downside). I tried to model something better a few years ago, but my results were inconclusive. 35/15 is pretty good. What this stuff means for the Golden Butterfly's bands is anybody's guess!</pedantry>

[Greg]

Thanks Pet Hog. I appreciate if someone calls out bad math. My question still stands though of a potential on what makes sense for a rebalancing band in regards to a 20/20/20/20/20 (especially if two of the 20s are both stocks).

[Tyler]

Thanks Pet Hog. I appreciate if someone calls out bad math. My question still stands though of a potential on what makes sense for a rebalancing band in regards to a 20/20/20/20/20 (especially if two of the 20s are both stocks).

For a nice round number, I think 30/10 bands would work fine.

But practically speaking, I have found rebalancing due to hitting a band to be rare for two reasons:

1) Add new money to the lagging asset, and you naturally rebalance with no tax consequences.
2) I also like to sell just enough of the leading asset as I can every year to stay rebalanced while remaining in a favorable tax bracket. (I think Sophie has called that tax gain harvesting).

YMMV, but another way to maximize investment income is to minimize taxes paid. Just something to think about.