Cash Drag

[Tyler]

Somebody please correct me if I'm wrong, but it's my understanding that PP rebalancing almost always (or maybe always always) happens by hitting a 35% band, not by hitting a 15% band.  Is this true?  Has anybody had to rebalance because an asset went down so bad that it hit the 15% band?

Throw cash withdrawals into the picture, and the answer is "absolutely".  Life sometimes requires liquidity. 

[ochotona]

I have now about 18% of my portfolio in those Schwab "cash-drag" portfolios, and I like that it has that much cash at this point in the market; and 5% IAU, to boot.

[Kbg]

There's actually a fairly straightforward explanation for that. If we start at exactly 4x25% and the other three assets remain constant, an asset has to go up in price by a factor of 1.62 to trigger the 35% rebalance band. By constrast, an asset has to go down in price by a larger factor of 1.89 to trigger the 15% rebalance band. So in general, the latter should occur less often.

That implies the rebalancing bands are not symmetrical in terms of risk to reward.  Another flaw that needs to be fixed?

For my 3x ETF PP VP I rebalance at .5x and 1.5x the target weight. I'm not sure if it is a "flaw" per se or there was any thought behind 15/35% other than +/- 10%, but I decided to go symmetrical. If the theory is no predicting, why weight differently? The one logical thing for having a bigger hurdle for the downside rebalance is things tend to go down harder/faster/more than they tend to go up...so maybe that was why???

[mathjak107]

think of cash as owning call options on assets at lower prices.

[barrett]

think of cash as owning call options on assets at lower prices.

Interesting perspective. I kinda like it.

Me too!

[LC475]

For my 3x ETF PP VP I rebalance at .5x and 1.5x the target weight. I'm not sure if it is a "flaw" per se or there was any thought behind 15/35% other than +/- 10%, but I decided to go symmetrical. If the theory is no predicting, why weight differently?

Umm... I hate to break your bubble, but.... you are doing the exact same thing as the PP.  Your rebalance bands are 12.5% (.5X25) and 37.5% (1.5X25).  That is, +/- 12.5%.

Look, your bands and the PP bands are symmetrical... from a certain point of view.  In the PP, if an asset is either +40% of the target weight or -40%, that's when the rebalance happens.  15 is 60% of 25 (that is, 25 has to go down 40% to reach 15) and 35 is 140% of 25 (25 has to go up 40% to reach 35).  For you bands, it's just 50% instead of 40%.  Plus or minus 40% seems like a nice, reasonable figure -- a big range, allows for a lot of volatility between the marks.

The issue that I've thought of before (I was going to start a thread on it back when I was new, but never did) is that the 10-point gap or 40% gap however you look at it between 15 and 25 is bigger, proportionally, than the 10 point gap between 25 and 35.

35/25 = 1.4
25/15 = 1.66

So in that sense, 15 and 25 are actually quite a bit further apart from each other than 25 and 35.

To equalize it, if you're a fan of the 35% upper and want to keep it the same, you'd change it to:

35/25 = 1.4
25/17.857... = 1.4

So the new rebalancing bands are 35% and 17.86%.
If instead you loved the 15% band, you'd make it 41-2/3% and 15%.

That's all looking at the asset in isolation, as if the 25s and 35s and 15s were solid quantities, like 25 dollars becoming 35 dollars or 15 dollars.  But, of course, that's not what's happening in the PP.  The 25s and 35s and 15s are percentages of a whole, just part of the larger portfolio.  And so as the asset price fluctuates, not only the numerator changes, the denominator also changes.

So if one wanted to make the bands proportionally even and equal with respect to rises and falls in the asset price (not to the percentage of the portfolio represented by the asset,  which is what the above bands do) then you'd make your bands 35% and 17.06%.  That way, you buy if your asset goes down about 62% with respect to the rest of the portfolio, and you sell if it has gone up by about 62% with respect to the rest of the portfolio.  Or 38.65% and 15%.  With that one, you'd be waiting until the asset went either up or down by about 89% with respect to the rest of the portfolio.

So, mathematically, them's the facts and there they be.  However, irrational numbers just look so much more complicated on the page than rational ones.  Or in speech.  Can you imagine explaining to a friend, or Harry Browne saying on his radio show: "It's a very simple plan: 25% each, and then you rebalance whenever one of them hits either 35% or 17.0648% of the total portfolio."  Doesn't have quite the same ring, does it?  Also not very easy to remember.  Even rounding to 17%; 17 just doesn't have the pizzang.  And if you're rounding anyway, why not round to 15?

[MachineGhost]

So, mathematically, them's the facts and there they be.  However, irrational numbers just look so much more complicated on the page than rational ones.  Or in speech.  Can you imagine explaining to a friend, or Harry Browne saying on his radio show: "It's a very simple plan: 25% each, and then you rebalance whenever one of them hits either 35% or 17.0648% of the total portfolio."  Doesn't have quite the same ring, does it?  Also not very easy to remember.  Even rounding to 17%; 17 just doesn't have the pizzang.  And if you're rounding anyway, why not round to 15?

Gesundheit!  This is the next logical tweak for the Risk Parity PP.  I come up with 17.857142857142857142857142857143% though! 

[LC475]

I come up with 17.857142857142857142857142857143% though! 

So did I:

To equalize it, if you're a fan of the 35% upper and want to keep it the same, you'd change it to:

35/25 = 1.4
25/17.857... = 1.4

So the new rebalancing bands are 35% and 17.86%.

...but that's only the first iteration of the calculation.  It's 17.06...  in the end.

But maybe you're just making a joke about rounding!  If so,