Why Does The Rebalancing Bonus Work?

One of the best parts about Twitter is that you can follow super smart people and just read their interactions. For instance, yesterday I was creeping on Jeremiah Lowin, Jake from Econompic and Corey Hoffstein. These are three super smart finance guys who were going back and forth about the rebalancing bonus which was made famous by William Bernstein.

I will let you assess their comments as you wish, but I have a lot to say about rebalancing because it's so central to everything I profess so here's my general thoughts on the idea of a "rebalancing bonus."

The basic idea of a rebalancing bonus is that you will rebalance a portfolio away from higher risk assets when they've gone up in value and rebalance towards lower performing assets. It is, in essence, a systematic way of selling high and buying low. For instance, when your 50/50 stock/bond portfolio grows into 60/40 you want to rebalance back to 50/50 and sell the stocks to rebalance back towards the lower risk and lower performing asset because now you're overexposed to the higher risk asset.¹ That's simple enough. But why does this generate what Bernstein calls a rebalancing bonus?

There's a lot of debate about why this works or whether it works at all. And in fact, the results are quite different in nominal and risk-adjusted terms. If we look at the data since 1940 a portfolio of 50/50 stocks/bonds will do slightly better on a risk-adjusted basis and does 1.5% worse per year when rebalanced.

There's a bunch of interesting elements within these results:

1. Why is there a risk adjusted rebalancing bonus? The nominal results are obvious - stocks tend to beat bonds so a portfolio that becomes more and more overweight stocks will tend to do better on average. The superior risk adjusted returns are a point of contention in financial circles, though. I think it's pretty simple - the Global Financial Asset Portfolio is a lot more stable than most people think across time. It is currently 46% stocks and 54% bonds, but has been pretty close to a 50/50 portfolio since 1960 when reliable data is first available. So it makes sense then, if this is "the market" portfolio that rebalancing ever closer to this portfolio will generate a more efficient type of return. In other words, rebalancing away from stocks when they grow disproportionately in your portfolio is another way of rebalancing back towards the market portfolio.

Now, many people are quick to argue that "you can't eat risk-adjusted returns." This is bull shit in my experience. You don't rebalance to beat the market. You rebalance to ensure you don't get scared out of the market. This is the crucial element here. Again, it's all about behavioral alpha.

As I recently wrote, the suboptimal strategy that you stick with is better than the optimal strategy that you bail on. That's because you'll achieve a higher total return from the suboptimal strategy than you will trying to chase what looks like the optimal strategy (all the while generating lower returns due to higher taxes, higher fees and performance chasing). So you get more return from a suboptimal strategy that you stick with rather than the strategy you're constantly bailing out of. And if the suboptimal strategy is the strategy that generates higher risk-adjusted returns then you're absolutely eating more of the nominal returns than you will in the strategy that looks optimal on paper, but leads to worse behavior and lower nominal returns.

The above data is a perfect example of this. If you showed these results to 100 investors my guess is that most of them would want the higher return portfolio. But the higher return portfolio comes with a lot more risk. In other words, the portfolio that you don't rebalance is the portfolio that will expose you to more and more equity market risk over time and that means you're exposing yourself to more and more behavioral risk. So those same 100 investors are more likely to bail on the higher risk portfolio when it performs poorly. And on average, the lower return portfolio will generate an average higher investor return because it's the strategy that you actually stick with over time.

In sum:

• You might get a rebalancing bonus in risk-adjusted terms by rebalancing away from high performing assets. But that will come at the expense of lower nominal returns.
• The cause of this bonus is due to a rebalancing process that generates a portfolio that is more consistent with the Global Financial Asset Portfolio.
• Rebalancing is the process of instilling discipline in your portfolio by maintaining a consistently tolerable risk profile over time.
• Although the risk- adjusted rebalancing bonus is not guaranteed the odds of achieving behavioral alpha (better performance through better behavior) is increased because rebalancing results in a portfolio that you're more likely to stick with over time.

¹ - I'm an advocate of doing something a bit different than this. Countercyclical Indexing is a low cost and tax efficient form of indexing that rebalances periodically in order to account for the amount of risk that the markets expose to your portfolio. So, for instance, when 50/50 grows into 60/40 it might not be wise to rebalance back to 50/50 because your new 50% equity allocation might be riskier than your original 50% equity holding. So, late in a market cycle it typically makes sense to rebalance further away from stocks to something like 40% stocks. In doing so you not only benefit from the behavioral alpha of rebalancing consistently, but you also reduce the potential exposure to the equity markets when they are likely to be most volatile later in a market cycle.

Comments

[Augustus]

Several years ago I read a pretty good article which analyzed rebalancing methods. As you show, never rebalancing wins on the Returns comparison simply because equities have outperformed bonds over the study period. And you make note that the final result hides many periods of intense pain from the equity drawdown periods.

The recommended method from the paper was rebalance when the largest allocation became 20% overweight. Then rebalance only half way back to the target, somewhat letting the winners continue to run. The original 50/50 would rebalance when it became 60/40 and be reset to 55/45. No consideration of rebalance dates. Sorry I do not have a citation for the research.

Your comment about changing targets based upon an evaluation of which part of the economic cycle is identified does somewhat require the crystal ball. Who really knows how long an expansion can last. More interesting would be to know how the different market psychological evaluations are changing, probably more important than next year's earnings. Expansion of earnings multiples are the source of most gains over a two or three year period.

[sleek]

Cullen,

What about taxes?

Part of tax-planning is about deferring taxes to future periods when income is lower. Triggering capital gains by selling stocks at other times can act as a drag on portfolio returns.

[Cullen Roche]

Good question. It's all about the trade-off. Incurring small taxable events is better than being in a strategy where you risk taking a huge capital loss because you miscalculated your risk profile.

I think you should only rebalance once every few years though. You need to be super patient and reduce the negative impact of taxes while also maintaining a profile that you can stick with.

[dmarilley]

I wouldn't ever assume the average manager is Peter Lynch. Rebalancing is an admission of ignorance. Also, I've never heard of continuous dynamic rebalancing, so that ideal is where an intelligent discussion begins before adding capital market microstructure and fee friction into the discussion. The subtitle should be "Making Allocation Foolproof ". But too many fools manage money.