Rebalancing, really?

[hiriskpaul]

Obvious example,

aren't all examples obvious with hindsight?

How can an example not involve hindsight?

[EssDeeAitch]

I am aware that there are proponents of rebalancing a portfolio (restated by a Patrick Connolly of Chase De Vere in this weeks IC on the Portfolio Clinic column), the principle being to "take profits from investments that have done well and reinvest the proceeds in ones that have done badly".

...............

Or in short, my personal answer to the thread title is "Yes, really rebalancing - but rebalancing as you've described it, no way!"

Gengulphus

Rebalancing as described by others and challenged by me in truth. Thanks for the comprehensive response, always illuminating.

[Itsallaguess]

I think there's a potential performance hit from rebalancing that some investors might well be willing to take, if it means that their lifetime investment journey is one they can take more comfortably...

This is something that's touched on in Gengulphus's post below, and which I agree with after taking similar actions on some holdings in the past -

For example, I do some rebalancing of my portfolio and I am a proponent of doing so in a rather weak sense - essentially, my belief is that if imbalances grow big enough, practically every investor will find themselves wanting to do it: with only very rare exceptions (if any), the issue isn't whether an investor will decide to rebalance, but how big an imbalance is needed to get them to the point of making that decision.

https://www.lemonfool.co.uk/viewtopic.php?f=8&t=14832#p181661

I think sometimes that too much emphasis is given to overall investment performance, at the detriment of trying to find a more balanced approach that we can all comfortably manage as individuals.

Finding that balance might often mean that some future performance level might be lower than it might have been otherwise, if different, more uncomfortable choices had been taken, but I think it needs recognising that to stay the course over a lifetime of investment, it requires a level of comfort that we're happy with, first and foremost, and that will be different for all individuals.

So I personally don't buy the full-on 'run-your-winners' argument - yes, it might well promote a higher performance over the years, but if a different approach delivers adequate performance, and does so at a level of investment-comfort that allows that investor to stay the course, then that's got to be a better result than not being in the race at all, in my view....

To be clear though, I also don't agree with a full-on 'sell-your-winners' argument either - but I think a level of trimming back, and capital-redeployment for comfort reasons, is a quite valid middle-ground. It leaves a level of capital in the original winner, through which to achieve any future growth, and allows a level of capital to be transferred elsewhere, to perhaps turn into tomorrows winner too....

Cheers,

Itsallaguess

[Lootman]

I personally don't buy the full-on 'run-your-winners' argument - yes, it might well promote a higher performance over the years, but if a different approach delivers adequate performance, and does so at a level of investment-comfort that allows that investor to stay the course, then that's got to be a better result than not being in the race at all, in my view....

To be clear though, I also don't agree with a full-on 'sell-your-winners' argument either - but I think a level of trimming back, and capital-redeployment for comfort reasons, is a quite valid middle-ground. It leaves a level of capital in the original winner, through which to achieve any future growth, and allows a level of capital to be transferred elsewhere, to perhaps turn into tomorrows winner too....

The mantra about running your winners is usually quoted alongside a plea to also "cut your losers". In other words it is the idea that if you restrict your losers to (say) a 10% fall each but leave your winners intact, then even if you pick fewer winners than losers you will still end up ahead because those winners will go up more than you lost on the losers.

And if you are judging your performance in terms of cap-weighted index funds, which many use as a benchmark, then those index funds effectively let their winners run as well. So if that is your strategy then what you are really doing is expecting an edge from cutting your losers, which index funds do not do.

Rebalancing, on the other hand, not only does not cut the losers but doubles down on them. So it really is the polar opposite.

In my case I do something different and cull my outliers, which means cutting back on my biggest winners but also dumping my real failures. I am not claiming a secret sauce there, merely an approach that is systematic, enables me to gradually simplify my portfolio and can be tax efficient. As I think you suggest, it's really as much about your personal goals, approach and risk appetite, rather than claiming any objective superiority.

[Bubblesofearth]

The only evidence I've ever seen from long-term studies of portfolio performance is that long-term buy and hold works better than any other approach at building wealth. There was a study done a while back looking at the performance of the original Dow Jones components and an investor would have done best if he/she had put an equal amount of money into each company and then done nothing. This approach out-performed both cap weighting and leaving alone and tracking the index.

Are there any studies that show rebalancing either builds wealth better than not rebalancing or reduces risk? I'm not talking about theoretical arguments (these can be made on both sides) but actual comparative studies of portfolio performance.

Also worth remembering that rebalancing isn't free, you will pay transactions costs.

BoE

[OLTB]

As I'm in the building phase of my various portfolios (HYP, passive and ITs) the way I am approaching rebalancing is rather than sell 'winning' assets to re-invest in those that have fallen behind, I top up the assets with dividends that have been generated from the existing investments and the odd bit of capital. I assume that the overall effect will be the same, but I am only having to deal with one set of dealing costs and am also able to run my winners. It will be interesting to see what the outcome eventually is (I do post the results each year, but as it was my first full year this year (apart from my HYP), there's not much of a record so far!).

Cheers, OLTB.

[argoal]

Are there any studies that show rebalancing either builds wealth better than not rebalancing or reduces risk? I'm not talking about theoretical arguments (these can be made on both sides) but actual comparative studies of portfolio performance.

I don't have a study, but using the Bogleheads backtesting spreadsheet (2016 version) I did a couple of simple portfolio backtests.

A 60/40 Total US Market / Intermediate Gov Bonds (1972-2015) gives the results for a rebalanced and non-rebalanced portfolio (excluding Costs).

Real returns for Rebalanced: CAGR 5.13% - Std Dev 11.56%
Real Return for non-Rebalanced: CAGR 5.07% - Std Dev 13.41%

Harry Browne Permanent Portfolio (25% Total US, 25% long term bonds, 25% short tern bonds, 25% Gold) (1985-2015)

Real returns for Rebalanced: CAGR 5.40% - Std Dev 6.65%
Real Return for non-Rebalanced: CAGR 5.63% - Std Dev 7.97%

There are obviously 100s of other portfolios I could have tested but these represent two very different approaches so may provide some insight.

The date ranges were chosen as they were convenient not to try to fit the result in any way. Different date ranges would probably provide different results.

The tests are sufficient to suggest to me that rebalancing gives a smoother ride without a significant cost in performance.

[Lootman]

The only evidence I've ever seen from long-term studies of portfolio performance is that long-term buy and hold works better than any other approach at building wealth. There was a study done a while back looking at the performance of the original Dow Jones components and an investor would have done best if he/she had put an equal amount of money into each company and then done nothing. This approach out-performed both cap weighting and leaving alone and tracking the index.

There are equal weight ETFs out there, so if you really believe that equal-weighted out-performs cap-weighted indices then it is easy to compare them. In practice equal-weight does well when smaller caps out-perform large-caps. In recent years the market has been led by Apple, Amazon etc. so equal-weight hasn't done so well.

I would not choose the Dow for any purpose since it is price-weighted, which makes no sense to me. It's also only 30 stocks.

Here is a chart of RSP (equal-weight S&P 500) versus SPY (cap-weighted S&P 500), showing slight under-performance over 5 years for equal weighting::

https://finance.yahoo.com/chart/RSP#eyJ ... 1dfQ%3D%3D

[hiriskpaul]

Recent equal weight paper by S&P: https://us.spindices.com/documents/rese ... ndices.pdf

I have just glanced through it, but the strategy does seem to have worked very well for the S&P 500, less so for other markets.

[Bubblesofearth]

I don't have a study, but using the Bogleheads backtesting spreadsheet (2016 version) I did a couple of simple portfolio backtests.

A 60/40 Total US Market / Intermediate Gov Bonds (1972-2015) gives the results for a rebalanced and non-rebalanced portfolio (excluding Costs).

Real returns for Rebalanced: CAGR 5.13% - Std Dev 11.56%
Real Return for non-Rebalanced: CAGR 5.07% - Std Dev 13.41%

Harry Browne Permanent Portfolio (25% Total US, 25% long term bonds, 25% short tern bonds, 25% Gold) (1985-2015)

Real returns for Rebalanced: CAGR 5.40% - Std Dev 6.65%
Real Return for non-Rebalanced: CAGR 5.63% - Std Dev 7.97%

There are obviously 100s of other portfolios I could have tested but these represent two very different approaches so may provide some insight.

The date ranges were chosen as they were convenient not to try to fit the result in any way. Different date ranges would probably provide different results.

The tests are sufficient to suggest to me that rebalancing gives a smoother ride without a significant cost in performance.

The problem here is that you have chosen a time period that basically covers the longest bull market in bonds in history. So rebalancing between bonds and equities is going to be massively flattered as a strategy compared to other time periods. Try the 40 or so years before the 70's for example.

BoE

[Bubblesofearth]

There are equal weight ETFs out there, so if you really believe that equal-weighted out-performs cap-weighted indices then it is easy to compare them. In practice equal-weight does well when smaller caps out-perform large-caps. In recent years the market has been led by Apple, Amazon etc. so equal-weight hasn't done so well.

Equal weight ETF's are completely different from equal weight on purchase long-term buy and hold portfolios. They basically churn holdings at regular intervals to maintain their equal weighting status.

BoE

[Lootman]

There are equal weight ETFs out there, so if you really believe that equal-weighted out-performs cap-weighted indices then it is easy to compare them. In practice equal-weight does well when smaller caps out-perform large-caps. In recent years the market has been led by Apple, Amazon etc. so equal-weight hasn't done so well.

Equal weight ETF's are completely different from equal weight on purchase long-term buy and hold portfolios. They basically churn holdings at regular intervals to maintain their equal weighting status.

I would not say "churn" there. There is rebalancing on a periodic basis to maintain the correct weightings. An equal-weight ETF will still out-perform cap-weighted if the equal weighting idea is sound.

Whether you should re-balance or not is a separate issue from whether you should equal weight or cap weight.

[argoal]

The problem here is that you have chosen a time period that basically covers the longest bull market in bonds in history. So rebalancing between bonds and equities is going to be massively flattered as a strategy compared to other time periods. Try the 40 or so years before the 70's for example.

BoE

I didn’t choose the periods. They were the periods for which I had data at hand. 45+ years is a pretty long data sample.

I seen to remember that in real terms the 70s were pretty brutal for bonds because of high inflation which meant investors lost money for a decade or more when adjusted for inflation.

The 70s were also no picnic for stocks and made the recent rather grandly named ‘Great Financial Crisis’ seem like a flesh wound in comparison.

I’m not sure that there is reliable financial data pre 1970s available as most of the indicies barely existed before then. If there is good data available then I’m sure that bogleheads would be very happy to include it in their tools.

[tjh290633]

The 70s were also no picnic for stocks and made the recent rather grandly named ‘Great Financial Crisis’ seem like a flesh wound in comparison.

I’m not sure that there is reliable financial data pre 1970s available as most of the indicies barely existed before then. If there is good data available then I’m sure that bogleheads would be very happy to include it in their tools.

The only one in regular use then was the FT Ordinary Index (FT30). That took a hefty blow in 1974 but recovered fairly quickly.

I was investing steadily at that time, and my portfolio value halved from end 1972 to end 1974, but had risen above the end 1972 level by end 1975. Income was little affected.

Here is my yield for the decade:

Year     Yield %
Dec-71     5.98%
Dec-72     4.57%
Dec-73     5.82%
Dec-74    11.46%
Dec-75     5.39%
Dec-76     6.40%
Dec-77     8.34%
Dec-78     5.41%
Dec-79     6.23%
Dec-80     5.90%

The 1974 effect is clear.

TJH

[Bubblesofearth]

I didn’t choose the periods. They were the periods for which I had data at hand. 45+ years is a pretty long data sample.

I seen to remember that in real terms the 70s were pretty brutal for bonds because of high inflation which meant investors lost money for a decade or more when adjusted for inflation.

The 70s were also no picnic for stocks and made the recent rather grandly named ‘Great Financial Crisis’ seem like a flesh wound in comparison.

I’m not sure that there is reliable financial data pre 1970s available as most of the indicies barely existed before then. If there is good data available then I’m sure that bogleheads would be very happy to include it in their tools.

If you look at figs 7&8 in the Barclays equity gilt study linked below you can see how the two asset classes have performed very well since the mid-70's. The high inflation period was exactly the right time to start investing in fixed interest basically because inflation and interest rates have been on a downward trend ever since,

Now take a look at the divergence in performance of stocks and bonds in the decades before the mid-70's. You really wold not have wanted to have been rebalancing from equities to bonds during that time.

https://www.courtiers.co.uk/news/barcla ... study-2016

[LooseCannon101]

Re-balancing sometimes works and sometimes doesn't. The only person who always makes a gain is the stockbroker or intermediary.

I wonder how much of business magazines' revenue comes from such intermediaries advertising their wares. Personally, I never sell any of my highly diversified world equity fund regardless of it's price.

[hiriskpaul]

There are equal weight ETFs out there, so if you really believe that equal-weighted out-performs cap-weighted indices then it is easy to compare them. In practice equal-weight does well when smaller caps out-perform large-caps. In recent years the market has been led by Apple, Amazon etc. so equal-weight hasn't done so well.

Equal weight ETF's are completely different from equal weight on purchase long-term buy and hold portfolios. They basically churn holdings at regular intervals to maintain their equal weighting status.

BoE

The benefits of rebalancing or not is what the thread is about. It is unfortunate that the paper I posted did not examine equal weighted buy and hold portfolios for the period. Nonetheless, the equal weighted S&P 500, with regular rebalancing, did beat the S&P 500 by 2.1% annualised over the 15 year period. About half of that they attribute to the small company factor, the rest to "anti-momentum".

They do offer an alternative way of looking at what is going on however. They consider the distribution of returns and come up with the well known observation that about two thirds of stocks underperform the average. They then look at the source of excess returns and find that "equally weighting among the alternative constituents offers the maximum expected participation in the relatively small number of outperforming stocks." I think it is a bit of a leap to claim that equal weight offers the "maximum expected participation", but it did appear to be superior to cap weighting. I think this adds weight to the argument in favour of equal weighting. Additional evidence towards equal weighting is the finding that the equal weighted S&P small and mid cap indices did not outperform the normal cap weighted ones by as much and that these smaller cap indices are less skewed towards the largest stocks. In other words, "the potential impact of equal weighting is likely to be greater in large-cap indices than in other capitalization ranges", effectively because the smaller cap indices are closer to equal weight than the large cap index.

To take that one step further, if you do not rebalance, then the portfolio will become skewed, i.e. no longer equal weighted, so from the argument above it seems to me that the longer a portfolio is left unbalanced, the less likely it is that it will gain "maximum expected participation in the relatively small number of outperforming stocks". If you take the view that equal weight is optimal, then you a need to rebalance back to it occasionally, or at least not allow it to get too skewed, otherwise you are likely to miss out on some of subsequent large gains by the minority of stocks.

[hiriskpaul]

The problem here is that you have chosen a time period that basically covers the longest bull market in bonds in history. So rebalancing between bonds and equities is going to be massively flattered as a strategy compared to other time periods. Try the 40 or so years before the 70's for example.

This is key to gaining a benefit from rebalancing. If you have a portfolio of securities that do equally well over a long period (or not too far apart), then you are very likely to gain from rebalancing. There are good theoretical reasons for this as well as observations. The difficulty is that you do not know in advance which securities will perform equally well!

[GeoffF100]

Here is another article about the S&P 500 Equal Weight Index:

https://www.forbes.com/sites/rickferri/ ... ffba793356

There were good times and less good times. The risk adjusted return appears to have been the same as for the market weighted index, before costs. Tracking the market weighted index is cheaper.

On this evidence, rebalancing a portfolio of individual shares is neither good nor bad, except for the costs. There may be practical reasons for trimming winners, but that is a different matter. I trim my largest individual share holdings to fill my CGT allowance, and buy other investments with the proceeds. I neither trim nor add to my losers. That approach has served me well.

Rebalancing between equities and bonds is good if the market jiggles around, but bad if it shoots off in one direction or the other, without tracing its steps. In particular, if the equities shrink and go on shrinking, eventually to nothing, rebalancing potentially causes you to lose your bonds as well as your equities. If you back test rebalancing on markets that always recovered it looks great.