re-Balancing your Portfolio
Once upon a time I was wandering through the Morningstar Conversations and ran across a statement by Larry Swedroe, commenting on a recent study. He said:
"What is likely to surprise most investors is that adding international assets to a portfolio during a period when they underperformed actually resulted in higher returns and lower risk"

>And you said, "Ya gotta be kidding," right?
Something like that. Anyway, if one rebalances periodically, selling the components of a diversified portfolio which did well in order to buy more of the components which did poorly, then, I thought, it's much like Dollar Cost Averaging. (See DCA stuff.)

>You buy more units at the low price with your investment dollars, eh?
Right. Remember this chart?

>Yeah. Stock A increases over ten years and stock B goes nowhere. It ends up ...
.. with zero gain. Right. Yet DCA-ing into stock B resulted in the greater portfolio gain because ...

>Because of the increase of stock price from $1.50 to $6.00, eh?
Stop interrupting! Anyway, it then seems eminently reasonable to consider buying both stocks, devoting a certain fraction of your portfolio to each and rebalancing periodically so that you're selling stock A and buying stock B and ..

>And doing this "periodically" is what reminds you of DCA. Am I right?
Pay attention. We're talking about diversification so instead of stocks A and B we're going to consider asset classes A and B, understand?

>Yeah. A is stocks and B is bonds.
Or A is Domestic investments and B is Foreign.
>Or A is Small Cap and ...
That's enough.

We consider two asset classes. We'll call them A and B. Whereas A increases linearly in price over a certain period, B decreases for the first half of the period then increases during the second half, ending up with a smaller annualized gain over, say, twenty years. The result is ...

>I know! You sell some of the good stuff and buy some of the lousy stuff and ...
You devote 30% to the lousy ... to the asset class B, rebalancing each year so as to maintain this percentage.

>Why rebalance once a year? Why not ... ?
That's what everybuddy does ... I think. However, if you do rebalance periodically (like quarterly, every six months or annually or whatever) you may be interested in this scenario:

  • We invest $8K per year, increasing 2% annually - cuz of inflation.
  • We rebalance every 2 months or every six months or every 12 months in order to maintain a particular split, like 75% stocks + 25% bonds.
  • We use a Monte Carlo simulation (see Monte Carlo) to see whether we can achieve a $1,000,000 portfolio, after thirty years.
  • We look at the Monte Carlo probability for each period and decide whether there's much difference between six month rebalancing or twelve month or ...

>Did you say Monte Carlo probability?
Yes. We run a thousand scenarios and calculate how often our final portfolio achieves the $1M figure. Since we're trying to predict the future, it's fiction, you know. The best we can do is see if there's a significant difference between rebalancing periods.

I don't think there's much difference, except that if you rebalance too often, you pay trading fees, commissions, etc. and that'll drag down your gains. Besides, every twelve months gives a slightly higher gain ... slightly.

Here a chart based upon Bernstein's data:

>I take it the charts don't include trading charges.
Right. But for a more strategic time to rebalance perhaps, you can take a peek at this Strategy or, if you really want to know when to rebalance, check this out :^)

>And why 25% or 30%? Is that some optimal percentage? Is that the best ...?
Of course not. I'm just giving you an example. I could just as easily ...

>And I assume that when you put money into the poorer performing sector you're expecting some dramatic future gains.
Quite true. Here's some other pictures which don't make such simple-minded price changes in the two asset classes:

>Where did these charts come from? Did you ...?
I generated them with a spreadsheet, the price changes in the two asset classes being generated randomly every time I pressed F9 to re-calculate and ...

>You haven't indicated how the shares change, how you decrease one asset and ...
Here's a picture of the spreadsheet:

Happy? Note that the Standard Deviation (or Volatility) is less than either asset class.
Right-Click on the picture, above, and "Save Target" to download the .ZIPd spreadsheet

>And why just two component? Why not lots more, like Domestic Large Cap Stocks and Fixed Income Mutual Funds and Cash and Foreign Large Cap and ...
I'll consider it. In the meantime, here's a chart which has three asset classes, including the EAFE Index (Europe, Australasia and Far East) and, although the annualized gain was smaller than U.S. Large Caps, the Volatility was reduced. In this chart we buy a bunch in 1978, rebalance yearly, and watch:

>And what about some of your mathematical bumpf? Can't you prove anything? Have you lost your formulas? Have you ...?
I said I'll consider it.

>Stop interrupting.

for Part II.