Shouldn't Rebalancing Result in Higher Returns?

[ejman]

From the samples I've checked it seems that periodically rebalancing a portfolio (say once a year as recommended) results in minimal gains at best in long term returns. This really seems counter intuitive. Shouldn't sell high buy low result in significant higher returns over time? I see that rebalancing does diminish volatitily significantly but why the limited effect on returns?

[haha]

Quote:

Originally Posted by ejman From the samples I've checked it seems that periodically rebalancing a portfolio (say once a year as recommended) results in minimal gains at best in long term returns. This really seems counter intuitive. Shouldn't sell high buy low result in significant higher returns over time? I see that rebalancing does diminish volatitily significantly but why the limited effect on returns?

Maybe because you truncate occasional huge gains like the mid and late 90s.

Ha

[jIMOh]

Depends on what you are buying and what you are selling.

Depends on the trigger for rebalancing-
are you using a date (once a year)
are you using a percentage (one asset class has 10% more than another)
are you using a lower percentage (one asset class has 5% more or less than another)

The more often you rebalance (date wise) the lower return difference you will see
The higher the percentage you use, the higher the returns you will see (with higher volatility expected as well).

What is YOUR goal when you rebalance?

[Rich_by_the_Bay]

It does seem counterintuitive but everything I have read that's credible says the same thing: rebalancing probably has little or no beneficial effect on long term returns. It's main benefit is to keep your asset allocation close to where you want it year after year.

What is intuitive to me is to realize that at any given time the market is more likely to go up than down historically, at least time. So if you sell stocks to rebalance, those stocks would have been more likely to rise than fall the next year. You might scrape of your gains from the prior year, but in doing so you forego some gains the next year.

[cute fuzzy bunny]

You're cutting off momentum, which tends to take more than the one year rebalance cycle to turn.

Now, if you only rebalance when one asset class is clearly overvalued and send the money into a clearly undervalued asset class, you might have something. If you could figure out that whole over/under valued thing.

I think I saw some study that said if you rebalanced every 2-3 years you'd have better results than doing it annually. Annual rebalancing actually dampened returns.

This is basically the same problematic mechanism people who hold individual stocks encounter when they 'sell the winners'. You lose all that momentum and end up with a portfolio of losers.

Of course, if you rebalanced your fat stock portfolio into bonds at the end of 1999 you became a very happy camper.

[ejman]

So, would it make sense if I have the (more or less standard) 60/40 to wait until its 80/20 (or 40/60) before rebalancing? Would returns actually be higher then?

I have been a good little boy and rebalancing once a year as adviced. I am now wondering if maybe I should sign up with procrastinators anonymous on this particular score.

[cute fuzzy bunny]

Well, logically in most cases your 'real' returns come from equities, not from bonds. Bonds are more of a volatility reducer most of the time. Granted with period of exception.

So if you waited for 80/20, yes you'd have pretty good returns. If you waited for 40/60...not so much...

In most peoples cases, rebalancing reduces volatility, it doesnt increase returns.

[mb]

Quote:

Originally Posted by cute fuzzy bunny I think I saw some study that said if you rebalanced every 2-3 years you'd have better results than doing it annually. Annual rebalancing actually dampened returns. .

I think that was in one of Bernstein's books. The idea was to match the time between rebalancing with the typical time for a bull market trend and therefore avoid missing the "momentum" that you mentioned.

[haha]

Quote:

Originally Posted by mb I think that was in one of Bernstein's books. The idea was to match the time between rebalancing with the typical time for a bull market trend and therefore avoid missing the "momentum" that you mentioned.

Not to quibble, but all this contradicts the EMH. The whole asset allocation and index fund paradigm is based on this foundation.

If EMH holds, you cannot simultaneously increase returns and decrease volatility. That is the financial analog of a perpetual motion machine.

But hey, when did logic ever affect what people believe?

Ha

[mb]

Quote:

Originally Posted by haha Not to quibble, but all this contradicts the EMH. The whole asset allocation and index fund paradigm is based on this foundation. If EMH holds, you cannot simultaneously increase returns and decrease volatility. That is the financial analog of a perpetual motion machine. But hey, when did logic ever affect what people believe? Ha

You are correct but IRRC you are not a "true believer in the EMH gospel" anyway.

[haha]

Quote:

Originally Posted by mb You are correct but IRRC you are not a "true believer in the EMH gospel" anyway.

Good spot! You do recall correctly, but I didn't see a reason to include that in the post- since whether I believe it or not hasn't much to do with the contradiction.

Ha

[samclem]

Right, rebalancing doesn't do much , or maybe anything, to improve returns over the long run. It does help reduce volatility. And, THAT can produce higher returns. "What?"

Here's how I think of it.

In general, asset classes that have higher volatility (e.g. small emerging market stocks) provide higher returns than asset classes with lower volatility (e.g large US blue chip stocks). If they didn't, nobody would invest in them.

But, these high-return asset classes are risky-they can take a dive for years and lose large portions of their value. Yes, over a very long time they will provide higher returns, if you bet on one there's a good chance you'd go broke waiting for the higher return.

But, if you select a large number of these risky (high return) assets and choose them carefully so their returns have low or even negative correlations, you can enjoy the benefits of all the high returns without the big swings in the total value of your portfolio. If it were possible (it isn't) to do this perfectly, an investor could achieve very low volatility in overall portfolio value while holding nothing but high-risk investments. He'd get all the benefit of the high returns without the gut-wrenching swings in overall portfolio value. But, you have to rebalance or else during a prolonged cycle or you'll wind up making bets on just a few sectors--producing a high risk overall portfolio with no commensurate gain in overall returns. Rebalancing lets those of us who believe in the importance of asset allocation to take the risk that is key to higher overall returns. So, in a roundabout way, rebalancing does increase portfolio performance.

Yes I've heard that a two to three year rebalancing period improves returns compared to rebalancing every year. For the reasons above, I'd argue for rebalancing when your percentages get out of whack.

[samclem]

Quote:

Originally Posted by haha If EMH holds, you cannot simultaneously increase returns and decrease volatility. That is the financial analog of a perpetual motion machine.

True only when considering one asset class at a time. Not true when considering an overall portfolio. Each asset class could behave according to the EMH, but the portfolio could provide the same return as the average of the asset classes with much less volatility than any single component.

[jIMOh]

Quote:

Originally Posted by cute fuzzy bunny You're cutting off momentum, which tends to take more than the one year rebalance cycle to turn. Now, if you only rebalance when one asset class is clearly overvalued and send the money into a clearly undervalued asset class, you might have something. If you could figure out that whole over/under valued thing. I think I saw some study that said if you rebalanced every 2-3 years you'd have better results than doing it annually. Annual rebalancing actually dampened returns. This is basically the same problematic mechanism people who hold individual stocks encounter when they 'sell the winners'. You lose all that momentum and end up with a portfolio of losers. Of course, if you rebalanced your fat stock portfolio into bonds at the end of 1999 you became a very happy camper.

Smart Money magazine did an article on this years ago (1998? 1999? it was before the tech crash). It stated the same thing- you need to let winners ride. Then set a higher percentage as the rebalance trigger (5% too low, think 10% or 15%. If you see 20% gains, it is OK to sell some of them type thinking).


--
The way I do it is this:
1) I rebalance twice a year
a) I contribute to 401k 2X month
b) I contribute to my Roth Jan-August @$625/mo
c) wife's Roth is a sector portfolio

2) the first rebalance is in June. I do two things
a)sell 1% of all holdings into bonds (this is a gradual move from 100% equities to 90/10 or 80/20 over a 5 year period). I only sell if I am at a profit of more than 1% since January (for example sold in June of 2007 but did not sell in June of 2008)
b) adjust deposit amounts so the lowest performing assets of the year get most of the deposits. I might adjust my Roth deposits more month-month if I see things happening sooner (June-July-Aug is not a large enough deposit to rebalance)

3) the second rebalance is the last day of the year (Dec 29-30-31). I buy/sell for move in line with target allocations and also bump the bond percentage of contributions up 1% from previous year.

**Because I adjust contributions in June, I have only had to actually sell securities ONCE in 3 years following this practice.**

My 1% move every 6 months to bonds does complicate this some- but I understand it.

4) wife's sector allocation in her Roth is more of a weighting strategy- we have 7 sectors we buy each month for 12 months. Once sector is always overweighted (right now it is financials). Generally accumulating the sector which is getting hit hard at that time and maintaining positions in the others.

I'll let you know how this turns out- we just started doing the sector strategy in July.

[haha]

Quote:

Originally Posted by samclem True only when considering one asset class at a time. Not true when considering an overall portfolio. Each asset class could behave according to the EMH, but the portfolio could provide the same return as the average of the asset classes with much less volatility than any single component.

Perhaps. But that was not the OP's question.

[samclem]

Quote:

Originally Posted by haha Perhaps. But that was not the OP's question.

Sorry. I was responding to your post. That's probably what got me steered away from the OP's question.

[haha]

Quote:

Originally Posted by samclem Sorry. I was responding to your post. That's probably what got me steered away from the OP's question.

I appreciated your comment. If I were doing this kind of investment I would consider this point very important, and really try to understand the implications of it.

Ha

[ejman]

Quote:

Originally Posted by haha Not to quibble, but all this contradicts the EMH. The whole asset allocation and index fund paradigm is based on this foundation. If EMH holds, you cannot simultaneously increase returns and decrease volatility. That is the financial analog of a perpetual motion machine. But hey, when did logic ever affect what people believe? Ha

Hum, my noggin is starting to hurt. What I see is that say 60/40 portfolio that is rebalanced has about the same return as one that is not (over a long period of time) . So based on the quote above shouldnt the returns of the non rebalanced portfolio (the one with higher volatility) be actually higher than the rebalanced one? I guess this is the reverse of my original question.

[haha]

I don't think it is possible, according to the theory anyway, that there is some magic optimal ratio where your returns will be greater, and your risks lower.

However, this is not my game, and I don't really know the answer.

Ha

[audreyh1]

Quote:

Originally Posted by ejman From the samples I've checked it seems that periodically rebalancing a portfolio (say once a year as recommended) results in minimal gains at best in long term returns. This really seems counter intuitive. Shouldn't sell high buy low result in significant higher returns over time? I see that rebalancing does diminish volatitily significantly but why the limited effect on returns?

1. Because not all asset classes have the same long-term average return. Some are much higher than others.

2. The advantage of having a portfolio of several poorly correlated asset classes is lower overall volatility. The advantage of rebalancing such a portfolio is capturing a lot of the gain of an unbalanced portfolio while enjoying lower volatility.

If you don't care about volatility, for maximum long-term performance it's better to load up on small cap stocks or whichever is the asset class with the highest long term performance.

Audrey