Allan Roth: Boosting Returns With Rebalancing

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Random Walker
Posts: 2997
Joined: Fri Feb 23, 2007 8:21 pm

Allan Roth: Boosting Returns With Rebalancing

Post by Random Walker » Tue Mar 20, 2018 12:36 pm

http://www.etf.com/sections/index-inves ... ebalancing

I’ve always bought into this concept of possibly boosting returns by rebalancing between asset classes with similar expected returns, for example different equity classes. I realize that in an overall up market rebalancing from stocks to bonds will decrease returns and that overall rebalancing is primarily for risk control.

Dave

User avatar
Artsdoctor
Posts: 3454
Joined: Thu Jun 28, 2012 3:09 pm
Location: Los Angeles, CA

Re: Alan Roth: Boosting Returns With Rebalancing

Post by Artsdoctor » Tue Mar 20, 2018 1:00 pm

One would probably want to examine the data surrounding rebalancing, rather than generalizing.

https://personal.vanguard.com/pdf/ISGPORE.pdf

If you get a bit of a gain with your rebalancing, it's nice. But I don't think you can say that rebalancing will consistently "boost returns." At least the data would not consistently support that.

RadAudit
Posts: 3006
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Alan Roth: Boosting Returns With Rebalancing

Post by RadAudit » Tue Mar 20, 2018 1:05 pm

Roth says
So far in the first 18 years of the century, rebalancing has boosted returns.


I wonder how much the 18 year trend was helped by rebalancing just after 2008 - 2009? And around 2000, if I recall correctly.
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The Calvary isn't coming, kids. You are on your own.

User avatar
saltycaper
Posts: 2650
Joined: Thu Apr 24, 2014 8:47 pm
Location: The Tower

Re: Allan Roth: Boosting Returns With Rebalancing

Post by saltycaper » Tue Mar 20, 2018 1:10 pm

Taking more risk may or may not boost returns. Market timing may or may not boost returns.
Quod vitae sectabor iter?

azanon
Posts: 1886
Joined: Mon Nov 07, 2011 10:34 am
Location: Little Rock, AR

Re: Allan Roth: Boosting Returns With Rebalancing

Post by azanon » Tue Mar 20, 2018 2:15 pm

This is contrary to my understanding. It was my understanding that, generally speaking, rebalancing would, on average, lower both your risk and return since, again on average, you'll be selling stocks and buying bonds since stocks grow faster than bonds long-term.

User avatar
goingup
Posts: 3189
Joined: Tue Jan 26, 2010 1:02 pm

Re: Allan Roth: Boosting Returns With Rebalancing

Post by goingup » Tue Mar 20, 2018 2:18 pm

Allan Roth's article pretty much contradicts my understanding about rebalancing. The Vanguard article that Artsdoctor posted has this on page 4:

As expected, in most of our simulations (more than 70%), a nonrebalanced 50% stocks/50% bonds portfolio resulted in 30-year returns that were greater than those of an annually rebalanced 50%/50% portfolio. The median return of the drift portfolio was 7.6%, compared to the median rebalanced portfolio return of 7.1%.

In most cases rebalancing reduces returns because you're selling stocks to buy bonds. When stocks crater and you sell bonds to buy them, and then stocks recover--then you get a boost in returns!

azanon
Posts: 1886
Joined: Mon Nov 07, 2011 10:34 am
Location: Little Rock, AR

Re: Allan Roth: Boosting Returns With Rebalancing

Post by azanon » Tue Mar 20, 2018 2:29 pm

Ok i missed this when i skimmed it earlier: He intentionally chose a measured period where stocks and bonds performed similarly. So yeah sure if you pick a period where the actual returns are the same, then rebalancing should raise returns. That's just basic "free lunch" of a diversified portfolio.

But then he goes on to (indirectly) admit that if stocks do outpace bonds over a long period of time, the non-rebalanced portfolio (say one starting at 50% going to 70% stock) would outperform a rebalanced 50% portfolio.

maj
Posts: 412
Joined: Wed Jun 25, 2008 2:58 pm

Re: Allan Roth: Boosting Returns With Rebalancing

Post by maj » Tue Apr 10, 2018 2:08 pm

It is important to include not only gains and losses in the comparison between rebalanced and buy&hold, but also the emotional tolerance of the investor who may not wish to watch a 70% stock holding fall to 35% in a bear market.

User avatar
nisiprius
Advisory Board
Posts: 36492
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Allan Roth: Boosting Returns With Rebalancing

Post by nisiprius » Tue Apr 10, 2018 2:37 pm

All I want to do here, having had too many discussions--it's one of those things like the Monty Hall problem that's very difficult to get agreement on--is to note that William J. Bernstein has said that if the assets follow a random walk, then there is no rebalancing bonus. The rebalancing bonus only occurs if a) the assets show mean reversion, and b) the rebalancing interval resonates with the period over which mean reversion is observed.

If you believe (and some people do) that rebalancing can manufacture extra return from pure volatility, without assuming mean reversion in the assets, then your mental model of rebalancing is incorrect.

Oh, one more point. It is very seductive when thinking e.g. about rebalancing during 2008-2009, to imagine the rebalancing rule as "firing" somewhere close to the bottom. In reality, what happens in the the rule "fires" several times on the way down, and several times on the way up, and the gains from rebalancing do not correspond to the distance between the top and the bottom. They correspond to what happens when you do your best to pair up sales events and purchase events, and the result of rebalancing tends to be small, not always positive, and depend on the considerable luck of how the rule happens to interact with the particular history of how the prices moved.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
neurosphere
Posts: 2936
Joined: Sun Jan 17, 2010 1:55 pm

Re: Allan Roth: Boosting Returns With Rebalancing

Post by neurosphere » Tue Apr 10, 2018 3:51 pm

I have evolved in the way I think about rebalancing (btw I haven't read the article). I've never put these thoughts down in words before, but let me give it a first-pass effort here, and revise if necessary.

First my comparison for rebalancing is not to "not rebalancing". "Nobody" "not rebalances". :D That is, suppose a young investor starts out at 90/10 stocks/bonds. Assuming a long life, she probably won't end up with that allocation, and will want/need more bonds at 90 years old. That is, I accept the notion of a glide path as a default. And we can discuss/define terms all day, but I consider a glide path one type of rebalancing. For example, the glide path one sets for themselves is based on RISK tolerance at each stage of life. And we can "rebalance" around that glide path by any frequency we want.

Can we agree that if we rebalance daily (e.g. microchanges to our portfolio so that we maintain our desired risk/glide path), there is no concept of a rebalancing "bonus" here?

But if we rebalance less frequently, perhaps we can capture some bonus, without increasing risk? Rather than gliding smoothly, perhaps we decrease our stock allocation every 5 years. For example, if we wish to glide from 100% stocks to 0%. In year 2 we might be at 99% stocks. But what if we wait until year 5 to jump down to 95% stocks? In this case, we've increased risk (and possibly return) because on average our stock allocation is higher. But what if in year 2.5 we go from 100% stocks to 95% stocks. Then in year 7.5 go to 90%...etc. That's half the time we are above a smooth path, and half the time we are below. On average, our risk is similar (essentially oscillating about the daily rebalancing line).

Does this kind of strategy get us a bonus? I'm not smart enough to know the answer, and it requires knowledge of things like momentum, or reversion to the mean and other fancy words for which my education and intelligence to not allow me to completely understand. :confused
If you have to ask "Is a Target Date fund right for me?", the answer is "Yes".

dbr
Posts: 27207
Joined: Sun Mar 04, 2007 9:50 am

Re: Allan Roth: Boosting Returns With Rebalancing

Post by dbr » Tue Apr 10, 2018 4:51 pm

azanon wrote:
Tue Mar 20, 2018 2:29 pm
Ok i missed this when i skimmed it earlier: He intentionally chose a measured period where stocks and bonds performed similarly. So yeah sure if you pick a period where the actual returns are the same, then rebalancing should raise returns. That's just basic "free lunch" of a diversified portfolio.

But then he goes on to (indirectly) admit that if stocks do outpace bonds over a long period of time, the non-rebalanced portfolio (say one starting at 50% going to 70% stock) would outperform a rebalanced 50% portfolio.
Yes, but did you read that the rebalanced portfolio he compares to is 60% stocks on the grounds that the buy and hold portfolio averages 60% stocks and he understands that when stocks outperform bonds that obviously higher allocations to stocks outperform lower allocations to stocks. I actually can't tell exactly what he is comparing to what or what the recommendation really is. Allan Roth is basically one of the good guys but this particular article is probably not a best effort.

longinvest
Posts: 2907
Joined: Sat Aug 11, 2012 8:44 am

Re: Allan Roth: Boosting Returns With Rebalancing

Post by longinvest » Tue Apr 10, 2018 11:19 pm

dbr wrote:
Tue Apr 10, 2018 4:51 pm
azanon wrote:
Tue Mar 20, 2018 2:29 pm
Ok i missed this when i skimmed it earlier: He intentionally chose a measured period where stocks and bonds performed similarly. So yeah sure if you pick a period where the actual returns are the same, then rebalancing should raise returns. That's just basic "free lunch" of a diversified portfolio.

But then he goes on to (indirectly) admit that if stocks do outpace bonds over a long period of time, the non-rebalanced portfolio (say one starting at 50% going to 70% stock) would outperform a rebalanced 50% portfolio.
Yes, but did you read that the rebalanced portfolio he compares to is 60% stocks on the grounds that the buy and hold portfolio averages 60% stocks and he understands that when stocks outperform bonds that obviously higher allocations to stocks outperform lower allocations to stocks. I actually can't tell exactly what he is comparing to what or what the recommendation really is. Allan Roth is basically one of the good guys but this particular article is probably not a best effort.
I think that Allan Roth is saying that it's only fair to compare the performance of portfolios with similar average AA. Comparing portfolios with differing average AA would be unfair.

Let me illustrate this with an extreme example: it's pretty obvious that comparing the performance of a 100% bonds portfolio with that of a 100% stocks portfolio is unfair, because both portfolios have widely differing volatility. The same principle applies when comparing a rebalanced portfolio and a buy-and-hold one. It would be unfair to use the initial or the final AA of the buy-and-hold portfolio for comparison with a rebalanced portfolio; one should pick the average AA of the buy-and-hold portfolio.

I've explained this in another topic, recently. I was criticising a Vanguard paper which did an unfair comparison.

Rebalancing is profitable?
longinvest wrote:
Sun Apr 01, 2018 8:11 am
JustinR wrote:
Sun Apr 01, 2018 6:06 am
Check out Vanguard's white paper of rebalancing: https://www.vanguard.com/pdf/ISGPORE.pdf

The very first thing it says is:
The primary goal of a rebalancing strategy is to minimize risk relative to a target asset allocation, rather than to maximize returns

Then they find that never rebalancing will give you the highest return:
If a portfolio is never rebalanced, it tends to gradually drift from its target asset allocation as the weight of higher-return, higher-risk assets increases. Compared with the target allocation, the portfolio’s expected return increases, as does its vulnerability to deviations from the return
of the target asset allocation. To illustrate this, we compared two hypothetical portfolios, each with a target asset allocation of 50% global stocks/50% global bonds for the period 1926 through 2014; the first portfolio was rebalanced annually, and the second portfolio was never rebalanced. As the figure shows, and consistent with the risk-premium theory, the never-rebalanced portfolio’s stock allocation gradually drifted upward, to a maximum of 97%, and was 81% on average for the period. As the never-rebalanced portfolio’s equity exposure increased, the portfolio displayed higher risk (a standard deviation of 13.2%, versus 9.9% for the rebalanced portfolio) and a higher average annualized return (8.9% versus 8.1%, respectively).
The thing is that the portfolio's asset allocation was 81/19 stocks/bonds on average (see the part I put in red above). It shouldn't be compared to a rebalanced 50/50 stocks/bonds portfolio, but to a rebalanced 81/19 stocks/bonds portfolio.

Let me try to do a fair comparison using Portfolio Visualizer, using 2000 as the end date to maximize the higher return of stocks. I'll use two asset classes for which we have historical returns since 1972: US Stock Market and 10-year Treasury. (Total Bond Market return history starts in 1987).

Here are the returns and allocation drift of a 50/50 stocks/bonds portfolio which was never rebalanced:

Source: Portfolio VIsualizer
Image
Image

We see that stocks drifted from 50% to 78%. I'll estimate the average stock allocation as ((50% + 78%) / 2) = 64%.

Here are the returns of a 64/36 stocks/bonds portfolio which was rebalanced monthly:

Source: Portfolio Visualizer
Image
We see that both portfolios had similar returns, overall. I see no surprise, there.

The point is that an investor willing to let his portfolio drift from 50/50 to 78/22 stocks/bonds is actually investing into a portfolio which averages 64/36 stocks/bonds. He would be best to invest into a 64/36 stocks/bonds portfolio from the start and stay the course by rebalancing his portfolio. Starting with 50/50 and letting the portfolio drift leaves him with a 78/22 stocks/bonds portfolio in 2000, just before a big stock market drop!
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VLB/ZRR

Post Reply