Does Rebalancing Really Pay Off?

[Onward]

Over the years I have seen wildly conflicting data on this subject. This study confirms what I have come to believe: rebalancing may help you sleep, but it doesn't increase your returns much, and it could cost you a lot of $$.

(He doesn't mention commissions, taxes, spread, or the encouragement to "fiddle," but I'd argue these are further disadvantages to rebalancing.)

In summary, while rebalancing beat buy-and-hold most of the time, when it did beat buy-and-hold it was by a small amount, while when buy-and-hold beat rebalancing it was by about twice as large an amount. Between these two effects, the mean result was essentially a wash.

Does Rebalancing Really Pay Off??

 

[easysurfer]

Over the years I have seen wildly conflicting data on this subject. This study confirms what I have come to believe: rebalancing may help you sleep, but it doesn't increase your returns much, and it could cost you a lot of $$.

(He doesn't mention commissions, taxes, spread, or the encouragement to "fiddle," but I'd argue these are further disadvantages to rebalancing.)




Does Rebalancing Really Pay Off??

I think rebalancing helps you not panic so it protects you from yourself in times of fear and greed .

Plus, rebalancing does help take to emotion out by forcing you to add a more to assets that have gone down, and sell those that have gone up.

 

[Midpack]

Versus what? Buy and hold? Active investing?

The primary purpose of rebalancing is to maintain a chosen target asset allocation and the risk/returns associated with that AA, not to maximize returns. Obviously one could never rebalance and let equity % drift upward long term, and increase returns - but the volality/risk is greater as well. But corrections/downturns might be unnerving...

 

[MichaelB]

The paper seems a bit overly focused on William Bernstein and how he looked at rebalancing.

One of the concluding paragraphs

The results were surprisingly consistent across all portfolio pairs and strategies. The rebalanced portfolios did in fact beat the buy-and-hold portfolios most of the time. On average, a rebalancing strategy beat buy-and-hold about 70% of the time. However, on the occasions when buy-and-hold beat rebalancing, it was by an average margin about twice as great as the margin by which rebalancing beat buy-and-hold the rest of the time.

The author concludes that these two outcomes offset each other. If the focus portfolio survival and not total return, rebalancing may indeed be the better choice.

 

[FedExCourier]

Rebalancing is the way to sell high and buy low. Actually, Bernstein suggests rebalancing every couple of years or so. He believes in momentum carrying over in the market. So maybe that is kind of compromise in the two methods?

 

[LOL!]

In the limit, rebalancing is just market timing. And we know all about market timing. So folks generally write that rebalancing is to control risk and not to increase performance.

My feeling is that the best methods or ways to decide when to rebalance or market time have not been invented yet.

 

[NW-Bound]

...My feeling is that the best methods or ways to decide when to rebalance or market time have not been invented yet.

+1

There's no certainty in life, other than it will end at some point.

Take a little chance
It don't mean nothing
Sometimes you win, sometimes you lose

Take a little chance
It don't mean nothing
Sometimes you win, sometimes you lose

And don't you know that life is just a game
That you play, win or lose, it's all the same
​

 

[samclem]

Versus what? Buy and hold? Active investing?

The primary purpose of rebalancing is to maintain a chosen target asset allocation and the risk/returns associated with that AA, not to maximize returns. Obviously one could never rebalance and let equity % drift upward long term, and increase returns - but the volality/risk is greater as well. But corrections/downturns might be unnerving...

+1. And, to amplify MichaelB's point, I think most folks here are interested in making sure their investments have a very high probability of growing faster than inflation.

Without rebalancing, a portfolio will drift to higher and higher allocation to assets with the highest "risk" (because these also have the highest returns, so they'll gradually take over the portfolio). If this higher allocation to equities is acceptable, then just start off that way and you'll get an even higher expected average return. But, for must of us, maximizing the average expected return is NOT the goal (if it were, we'd invest in just a single stock, because that would have a higher average expected return than a broad index MF or ETF after expenses. It would also have a higher chance of ending with a zero balance or failing to keep up with inflation).

In the limit, rebalancing is just market timing. And we know all about market timing.

I disagree. Would we say that Target Date funds, that rebalance their holdings every day to keep them at the stated asset allocations, are trying to time the market (i.e. pick peaks and valleys, buy and sell at advantageous times)? It's no different if an individual investor is rebalancing to a fixed allocation on a calendar basis or when assets get out of whack according to his established rebalancing %ages. The goal is to reset the allocation back to the target so risk can be controlled.

 

[Lsbcal]

I did not rebalance in 2009 until the SP500 was in an uptrend. I did not sell equities but was down from a 55% position in 2008 to about 43% in mid-2009.

This works for me because my nightmare is to go through a period like late 1929 to 1932. The market was violently up-down over many months. It was like nothing we've seen in our modern day lives. I could provide some data if people need to be convinced.

 

[NW-Bound]

What you said about "smoothing" out the return by rebalancing is valid.

What is not correct is the following:

... if it were, we'd invest in just a single stock, because that would have a higher average expected return than a broad index MF or ETF after expenses. It would also have a higher chance of ending with a zero balance or failing to keep up with inflation).

Because a single stock has both chances of making you rich as well as homeless, the expected return is not higher than that of the index. What is higher is the standard deviation of the return, i.e. risk. This is more like buying a lottery ticket.

A trade-off more worthy of consideration is whether one goes 50/50 in AA or 100% stocks like what some people here do. We all know 100% stock gives a higher long-term return, but with a higher up/down. If one can stand such volatility with a strong stomach, shouldn't he go for it?

The answer is no. Not when he is making withdrawal from the portfolio. A 100% all-stock may do better than one with some fixed income, statistically that is. But tools like FIRECalc show that when one makes withdrawal, that changes the equation. Left alone, a 100% stock portfolio has been shown to bounce back. But if you sell low to buy food, ain't much left to bounce.

 

[Lsbcal]

...
Without rebalancing, a portfolio will drift to higher and higher allocation to assets with the highest "risk" (because these also have the highest returns, so they'll gradually take over the portfolio). If this higher allocation to equities is acceptable, then just start off that way and you'll get an even higher expected average return. ...

That is my approach in blue. I just pick my max equity and do not let it get more then 1% higher. Requires a sell maybe 3 or 4 times per year in a good year. The opposite of the "dry powder" approach.

 

[Chuckanut]

The answer is no. Not when he is making withdrawal from the portfolio. A 100% all-stock may do better than one with some fixed income, statistically that is. But tools like FIRECalc show that when one makes withdrawal, that changes the equation. Left alone, a 100% stock portfolio has been shown to bounce back. But if you sell low to buy food, ain't much left to bounce.

+1

Well said. These things can't be looked at in isolation but as part of a life-plan. If buy-and-hold fits better into one's life plan, then do it. For me, balancing seems to work best as I project (guess) out into the future.

 

[LOL!]

I read the linked Edesess article. He wrote:

For each two-asset portfolio and each 360-month period, returns were calculated for a buy-and-hold strategy and for six rebalancing strategies in which the portfolios were rebalanced monthly, quarterly, semi-annually, annually, every three years and every five years.

OK, so he shows that rebalancing based on a calendar does not work. So what? We already knew that. Earlier, he wrote about using the calendar, but also about using rebalancing bands:

The rebalancing discipline can be applied at regular time intervals, such as annually or quarterly or it could be applied when a threshold of imbalance had been reached, such as when either stocks or bonds crossed the 2% “out-of-whack” threshold (the 52% threshold in the example).

And he mentioned the nemesis of rebalancing and reversion-to-the-mean, momentum.

Other studies have shown that for shorter time intervals, securities prices show signs of mean-reversion’s opposite: momentum. That is, a high return over a time interval has increased the likelihood that the return will be higher than average over the next interval.

I suggest that a change in momentum coupled with rebalancing bands would be a good time to perform any rebalancing acts on a portfolio. Show me the studies where a change in momentum is used in deciding if there is a rebalancing bonus.

Bonus link: Daryanani's paper on opportunistic rebalancing has been linked before. It does mention momentum: http://www.tdainstitutional.com/pdf/opportunistic_rebalancing_jfp2007_daryanani.pdf

Here is Daryanani's conclusion:

We have shown significant advantages of opportunistic rebalancing (look frequently and rebalance only when you need to) over traditional annual or quarterly rebalancing. Opportunistic rebalancing goes beyond simply controlling risk; it also increases return benefits by capturing sporadic buy- low/sell-high opportunities. The benefits from opportunistic rebalancing far outweigh the costs associated with trading, taxes, and looking. Look frequently and trade less to maximize your rebalance benefits.

This is the kind of rebalancing that I try to do and report in the http://www.early-retirement.org/forums/f44/lol-s-market-timing-newsletter-57042.html thread.

 

[samclem]

What is not correct is the following:

Originally Posted by samclem
... if it were, we'd invest in just a single stock, because that would have a higher average expected return than a broad index MF or ETF after expenses. It would also have a higher chance of ending with a zero balance or failing to keep up with inflation).

Because a single stock has both chances of making you rich as well as homeless, the expected return is not higher than that of the index. What is higher is the standard deviation of the return, i.e. risk. This is more like buying a lottery ticket.

It is also true that the average return of a single random stock is higher (after paying costs) than the average return of all the stocks, if owned in an ETF or MF that has any costs at all.
It >is< like buying a single lottery ticket--but paying virtually no "house take", VS buying every lottery ticket sold, but paying a higher .05% vig. The single lottery ticket has a higher average return just due to the costs. This is required by the way we compute "average"--right?

 

[NW-Bound]

After the consideration of "friction" in holding an index fund vs. a single stock, you are correct.

For an advantage of as low as 0.04%/year (ETF expense ratio) in expected return, the single-stock investor risks a huge range of outcomes. Not at all a good trade, although mathematically the single stock has an infinitesimal advantage.

How to explain that in real life few people would go for broke? How to reconcile mathematical models with human behavior? This is easily explained by Bernoulli solution to the St. Petersburg paradox.

In layman's term, even if a bet has a high expected return, and perhaps even infinite, you do not want to take it if it also has a finite chance of bankrupting you.

Here's an example. You have $1M. I ask you to flip a coin. Head, you lose your $1M. Tail, I give you $9M for $10M total. The expected value of the bet is $5M, which is 5X what you have now. Do you take the bet?

See how hard it is to take the above gamble, let alone the 0.04%/yr advantage?

 

[Lsbcal]

How much importance do you assign to rebalancing versus AA and other investment ideas like diversification?

My top 50:
1) AA
2) buy-hold versus market timing of large parts of the portfolio
3) diversification
4) safety on the bond side
...
...
48) rebalancing
49) what they say in the Wall St. Journal

 

[Totoro]

Here's an example. You have $1M. I ask you to flip a coin. Head, you lose your $1M. Tail, I give you $9M for $10M total. The expected value of the bet is $5M, which is 5X what you have now. Do you take the bet?

See how hard it is to take the above gamble, let alone the 0.04%/yr advantage?

+1

The hedge fund industry has found a creative solution to this dilemma though. You gamble other people's money as follows:

• Heads, my investor loses $1M. I get $20k (2% management fee)
• Tails, my investor wins $9M, of which I get 20% performance fee, $180k plus the 20k.

The smarter ones go a step further: Create two funds of 1M each, each take opposite sides of the bet. That way I get a garantueed 220k. Then close down the losing fund, purge its records, and have a great track record. Get more funds, rinse & repeat.



Amazingly, it's legal.

 

[samclem]

See how hard it is to take the above gamble, let alone the 0.04%/yr advantage?

I agree 100%. But, still, the premise implied in the link at the OP is correct: Taking a highly risky and concentrated "bet" can be mathematically likely to result in a higher average (mean) portfolio value than a more diverse bet--whether the more diverse bet is due to periodic rebalancing or not.
A "newbie is likely to be attracted to an approach that promises a higher mean portfolio value. So, we have to be prepared to do as you have done--show that trying to maximize "average expected future portfolio value" could be accomplished by picking a single stock at random, but this goal has almost nothing to do with the real-life goals of most retirees. Instead, they likely want an extremely high likelihood of attaining at least a moderate level of growth above inflation (which can't be done by picking one stock at random). Understanding the difference between these two things is key to the discussion of rebalancing.

 

[NW-Bound]

Amazingly, it's legal.

It's no different than Las Vegas and state lotteries.

 

[samclem]

How much importance do you assign to rebalancing versus AA and other investment ideas like diversification?

My top 50:
1) AA
2) buy-hold versus market timing of large parts of the portfolio
3) diversification
4) safety on the bond side
...
...
48) rebalancing
49) what they say in the Wall St. Journal

If having an appropriate target AA is the top priority, (and I think it is), then I don't know how rebalancing (or whatever we call the means used to keep the actual portfolio aligned with the AA) could be very far down on the list.

 

[Lsbcal]

If having an appropriate target AA is the top priority, (and I think it is), then I don't know how rebalancing (or whatever we call the means used to keep the actual portfolio aligned with the AA) could be very far down on the list.

Well you got me on that one.

I guess I was thinking about major rebalancing as some tend towards e.g. market goes down a lot and there are major shifts to rebalance. My rebalance regime is so minor it seems to be way down the lists of concerns. It is designed to be small and incremental.

I recall someone saying that the smarty pants at Yale endowment rebalance daily (as apparantly Vanguard target date funds do). That is really an extreme form of incrementalism.

 

[NW-Bound]

When I made the posts earlier, I was responding to another post, and it was an interesting discussion but tangential to the topic of the OP. And the latter is about the effects of rebalancing vs. "buy and hold".


I have now read the article in the OP, where the author refuted some claims made by Bernstein. I am not familiar with Bernstein's study to see who is more correct. I also have not fully followed Michael Edesess's method to see if his conclusion is valid.

But I have these two comments to make.

1) In the phase of portfolio draw down, the WR may make a difference. It does not appear that either author considered that.

2) If the 2 authors are looking to maximize the expected return or average return, they are not doing what people here do when running FIRECalc. Most people want to be sure to stay afloat. They are willing to give up the potential for big gains, in return for an assurance of staying afloat. In engineering and optimal control theory, we talk about this "Minimax" problem, which seeks to have the best "worst case" result. And this is the same as the St. Petersburg Paradox I described earlier.

 

[Lsbcal]

...
But I have these two comments to make.

1) In the phase of portfolio draw down, the WR may make a difference. It does not appear that either author considered that.

2) If the 2 authors are looking to maximize the expected return or average return, they are not doing what people here do when running FIRECalc. Most people want to be sure to stay afloat. They are willing to give up the potential for big gains, in return for an assurance of staying afloat. In engineering and optimal control theory, we talk about this "Minimax" problem, which seeks to have the best "worst case" result. And this is the same as the St. Petersburg Paradox I described earlier.

These are very good points. My comments on them:
1) Withdrawals for spending very definitely are something that I had not really appreciated until it hit us in 2008-2009. Spending in a major down market makes the portfolio shrinkage accelerate as we know intellectually but seeing it happen to you may be a nightmare. We withdrew from FI and let the equity component slide.

2) For me the worst case as demonstrated in our history is 1929 to the bottom in 1932. I design my strategy with this one firmly in mind. Rebalancing into a down market is not for me because I know I could not stomach the ride. Frankly, a well designed market timing scheme could mitigate this but I know I'm in a tiny minority with this opinion.

 

[NW-Bound]

In real life there are a lot more factors that these studies cannot take into consideration. For example, even if rebalancing proves to be effective in a down market, a person drawing 5% WR will be more reluctant to pull the trigger compared to one drawing 2%. And the 1st person would not be wrong; he had a smaller margin for error if the market dropped more.

In matters like financing and investing, I tend to be more pragmatic and think more about people's psychology, my own included, rather than looking for some sure-fire methods. In my work, I could be a lot more theoretical and exact (one of my jobs had the title of "Scientist" because of the analytical work dealing with equations spanning pages) because I dealt with inanimate objects that could be tested and modeled very precisely. Understanding human emotions, including your own, is more difficult.

 

[clifp]

In layman's term, even if a bet has a high expected return, and perhaps even infinite, you do not want to take it if it also has a finite chance of bankrupting you.

Here's an example. You have $1M. I ask you to flip a coin. Head, you lose your $1M. Tail, I give you $9M for $10M total. The expected value of the bet is $5M, which is 5X what you have now. Do you take the bet?

See how hard it is to take the above gamble, let alone the 0.04%/yr advantage?

As PSA if anybody has opportunity to make a coin flip like this I'll happily risk upto $500K in return for $4.5 million. Just make sure the coin flip is done with your coin and not theirs