Does the rebalancing bonus exist and how can you earn it?

• Level: Advanced
• Reading time: 11 minutes

Rebalancing is a portfolio management technique that’s widely recommended to passive investors. You periodically sell your overperforming assets, buy more of your under-achievers, and reset your ETF portfolio’s allocation back to its target percentages. Like eating broccoli, it’s a wholesome practice that’s generally thought to be beneficial - especially when it comes to controlling portfolio risk. But just as exaggerated claims glom onto superfoods, somewhere along the way rebalancing got overhyped. Rebalancing, according to some, doesn't just keep your risk in check: it earns you extra return, too. This is the so-called rebalancing bonus. It’s a bold assertion, and there is some truth to it, but the rebalancing bonus is an elusive beast in reality. That being the case, let’s separate out the facts from the myths for ETF investors.

What rebalancing does to a conventional portfolio

Let’s look at a classic 60/40 portfolio: a world equity ETF and a government bond ETF. Over long periods, equities typically return more than bonds. Leave the portfolio alone and the equity share drifts upwards, from 60 % to 70 %, then 80 % and beyond - so long as asset returns conform to type. Rebalancing interrupts that drift. It repeatedly trims back equities and injects more capital into lower volatility bonds. Logically, that should cost you return over the long run, and historically it generally has. Here’s how that worked out for a 60/40 mix of world equities and euro government bonds, from 31 December 2006 to 30 June 2026: As you can see, annual rebalancing cost about 0.2 percentage points a year. Meanwhile, the pure buy-and-hold portfolio that’s never rebalanced drifted up to an 81:19 equity:bond share. Despite this, the risk-controlled 60/40 strategy only reduced annualised volatility by a minuscule 0.15 percentage points and suffered a worse deepest drawdown. Essentially, the rebalancing investor was unlucky. During the Global Financial Crisis (GFC), equities were the asset in sharp decline while bonds held up. The portfolio rebalanced back into equities in December 2008 just before stocks fell to rock bottom in February 2009. Equities were certainly on sale at the time and the purchase proved out. However, it meant the rebalanced portfolio recorded a slightly deeper loss than its buy-and-hold rival because it held more equities when the crisis peaked in February 2009. Since then equities have enjoyed relatively low volatility and haven’t been punished by another crash on the scale of the GFC. Meanwhile bonds were hit hard in 2022. Ultimately then, rebalancing has neither delivered a returns bonus nor meaningfully lowered risk for a conventional blend of World equities and intermediate euro government bonds over the past 20 years. Still, the outcome would look different if we ran the same comparison over the previous twenty years ending with the Global Financial Crisis. Much depends on where the big crashes fall on your personal journey and on the particular path markets take.

Where the bonus idea comes from: volatility drag

Yet the root of the rebalancing bonus is real and it starts with volatility drag. Suppose an ETF rises 50 % one year and falls 50 % the next. That amounts to a 0 % simple average annual return (the sum of yearly returns divided by the number of years). However, the ETF actually loses money because volatility (how much returns swing up and down) has eaten into compound growth. Here’s what happened: Year 1’s 50 % gain pumped €100 to €150. But year 2’s fall took a 50 % swipe from the now bigger €150 pot. €150 falls to €75. So our ETF gains €50 (nice) but loses €75 (Aargh!) from its higher balance. The bigger the swings, the greater the consequences for compound returns, so this ETF loses 13.4 % annualised over two years. Given the nature of the drag factor, the opportunity lies in selecting a diversified portfolio that is less volatile than its individual components. That cuts the volatility drag we’ve witnessed above - enabling a regularly rebalanced portfolio to potentially compound faster than the weighted average of its assets' compound growth rates. Researchers call this extra growth the diversification return. Willenbrock (2011) argues that the extra return is earned by rebalancing the portfolio.

Shannon's Demon: when rebalancing creates return

The pioneering information theorist Claude Shannon is credited with a thought experiment that shows how the diversification return works. Imagine an asset whose price either doubles or halves each period, with equal probability, like a coin flip. On average, the asset returns +25 % per period: which is the mean of +100 % and −50 %. (100 - 50) / 2 = 25. But the asset's compound growth rate is actually zero:

• 100 + 100 % = 200
• 200 - 50 % = 100

Buy this asset, hold it, and you go nowhere. Now put half your money in the asset and half in cash (earning 0 %). Then rebalance to 50/50 after every period: Each return cycle (100 % gain, 50 % loss) now leaves you with 1.50 × 0.75 = 1.125. That's a portfolio compound growth rate of about 6.1 % per period, from two ingredients that individually don’t grow at all. This rebalancing bonus - which seemingly conjures returns out of thin air - is known as Shannon’s Demon. It gets better when the returns of two such Demon assets are independent of each other. Rebalanced 50/50 each period, the portfolio doubles if both rise, halves if both fall, and grows 25 % if one rises and one falls. The compound growth rate works out at roughly 11.8 % per period. But there are no demonic forces at work. Instead, rebalancing systematically sells some of the asset that leapt ahead, and buys more of the one that fell behind. Because the moves are independent, and the identity of the winning/losing asset keeps switching, the portfolio reduces the volatility drag that undermines a buy-and-hold investor’s returns.

Shannon’s Demon in the real world

The UCITS ETF era has witnessed a real-life example of Shannon’s Demon, stemming from the interplay of gold and World equities during the past 20 years: With the help of Shannon’s Demon (or the diversification return) a 60/40 equities / gold portfolio (rebalanced annually) beat 100 % World equities by a considerable difference (1.19 percentage points per year). 100 % gold just pipped the diversified portfolio but banking everything on gold is certifiable. Moreover, it’s clear that the 40 % gold holding in the 60/40 isn’t enough by itself to close the gap between the 100 % equities and 100 % gold result. Shannon’s Demon made up the difference. We can see this as the rebalanced gold portfolio beat the buy-and-hold version – 10.17 % to 9.54 % annualised. The act of patient, systematic rebalancing cut volatility drag and increased return on the diversified portfolio.

What the demon needs

Pairs of assets that are more likely to earn a rebalancing bonus share these traits:

1. High volatility: The bonus grows with the square of volatility, so doubling volatility roughly quadruples it.
2. Low correlation: If both assets move together, there's nothing to sell high and buy low against.
3. Similar long-run growth rates: This reduces the chance of the rebalancing mechanism forcing you to keep feeding a structural loser.

The returns bonus can sometimes be earned from a conventional equities / bonds mix. But over the long run, bond ETFs usually have much lower volatility and expected returns than equities. So most of what rebalancing does there is give up equity growth in exchange for stability. Which leaves gold and commodities as the most likely partners for earning diversification return. Both gold and commodities have historically been highly volatile and have shown low correlation with equities over long periods.

Path dependency

We’ve already seen in the table above that a 60/40 equities/gold portfolio did earn a rebalancing bonus. (Let’s set aside for a moment the risky nature of that type of portfolio.) A 60/40 commodities portfolio did not earn a rebalancing bonus over the same twenty year period because commodities suffered a terrible bear market from 2008 through 2020. But if we reset the comparison to run from the end of 2017 (when commodities’ long losing streak had largely run its course), then commodities exhibit a rebalancing bonus too: We can see from the table that the annually rebalanced 60/40 equities/commodities portfolio beat its buy-and-hold twin by 0.56 % per year. That’s a healthy rebalancing bonus, albeit the comparison period is only eight and a half years.
Meanwhile Shannon’s Demon helped lift the same portfolio to within 0.5 % of the 100 % equities return (12.14 % to 12.63 % annualised). Intriguingly, return per risk (risk-adjusted return) is always higher for the rebalanced portfolios versus their buy-and-hold counterparts. In other words, rebalancing was effective in dampening risk without suppressing return during this period, though the margins were small. Finally, check out the well diversified portfolio (orange background colour) that splits the 40 % defensive allocation equally between gold, commodities and government bonds. This is the only portfolio that doesn’t involve placing a big bet on the future performance of a single diversifying asset. It turns out to have been a good call. Volatility is lower and return per risk higher than with 100 % equities. Yes, the gold-heavy portfolios were superior because the yellow metal was the winning asset class during this timeframe. But you can’t count on gold repeating the trick. It’s such a fickle asset. Still, the diversified portfolio did not earn a rebalancing bonus because bonds lost money during the period. Essentially, one weak asset in the mix cancelled out what the others gained from rebalancing. Which ultimately illustrates how flimsy the case is for the rebalancing bonus. Shannon’s Demon depends on: The growth-rate gap – The demon needs assets with similar long-run growth. If gold or commodities compound more slowly than equities over your holding period, the diversification return has to outrun that gap before rebalancing beats simply holding equities. Whether it has done so depends heavily on the period you pick. Unstable correlations – Low long-term correlations can hide extensive stretches where assets move together. The bonus shrinks whenever correlation rises. Trends hurt – In a long trend, rebalancing keeps selling the winner too early, and thus a drifting portfolio does better. Rebalancing pays most in choppy, mean-reverting markets. Costs – A bonus measured in fractions of a percent can be eroded by product costs and trading frictions. Taxes can also bite when rebalancing triggers realised gains. It's still mostly risk control – The main reason to hold gold or commodities alongside equities, bonds and money market ETFs is diversification. Any bonus is a welcome side effect, not a goal worth pursuing in itself. So rebalance because you want to control risk. If you also happen to hold assets that give the demon something to work with, treat any extra return as a bonus in the truest sense: welcome, but not guaranteed.