How to Rebalance Your ETF Portfolio

Say you picked 60% stocks and 40% bonds two years ago. Today it reads 68/32. Nothing broke; stocks went up more than bonds did, which is what they're supposed to do.

Drift changes how much risk you're carrying, and barely touches what the portfolio earns. So the decision worth getting right is the rule you'll follow for the next decade.

Drift decides how much risk you carry

McNamee, Paradise, and Bruno (2019) simulated a 60/40 portfolio for Vanguard from 1926 through 2018 and let one version run without ever being rebalanced. Its average equity allocation over the period was 85%. On paper the investor still owned a balanced portfolio; in practice they owned something close to an all-equity one, at 14.0% annualized volatility against 11.4% for the same portfolio rebalanced once a year.

Equities beat bonds over long stretches, so the riskier sleeve keeps growing into the space the safer one used to occupy. The allocation you chose was a statement about how bad a year you could sit through, and nothing tells you when you've drifted past it.

Rebalancing will not raise your return

The never-rebalanced portfolio in that study finished ahead on return: 8.74% a year against 8.19% for a version checked annually with a 5-point band. Both figures are after tax, assuming a 30% income rate and a 20% long-term capital gains rate, so some of that gap is just a portfolio that never sold anything and never realized a gain. The rest comes from the equity weight creeping toward 85%. Risk-adjusted, the ranking flips: Sharpe ratios of 0.46 for the drifting portfolio and 0.51 for the rebalanced one.

Change the window and even the return ordering reverses. Over 2005 to 2014, the same study found the portfolio left alone trailed a quarterly-rebalanced one by 5 percentage points after tax across the decade. It carried too much equity going into the 2008 correction and too little during the recovery, so it lost on the way down and again on the way up.

Drift pays for itself over 92 years. Over a single decade with a crash in it, it doesn't, and you don't get told in advance which one you're living through. Rebalancing keeps the portfolio recognizable while you wait to find out.

Any reasonable rebalancing rule works

Monthly or quarterly? A 5-point band or a 10-point one? The choice matters far less than it looks.

Vanguard tested those combinations across the same 92 years. Rebalancing monthly on any deviation at all took more than 1,100 separate rebalancing events; checking once a year and acting only past a 10-point band took 14. Both returned 8.20% a year, with a Sharpe ratio of 0.50. The difference shows up in risk: 63% average equity for the lazy version against 60% for the busy one. Vanguard's conclusion was that no one strategy dominates, and that picking a reasonable approach and sticking to it beats not rebalancing at all.

That is three points of average equity for 14 trades instead of more than 1,100. Both versions stay near the 60% target; the one left alone ended up at 85%. Check once a year on a date you'll remember, or whenever you're logged in for something else anyway, and act only when a band has been crossed.

A 5-point band is too wide for a small sleeve

The frequency is close to noise. The width of the band stops being noise once a sleeve is small.

Say you run a global three-fund portfolio, built from the cheapest broad fund in each sleeve: VT at 60% for world equities, BNDW at 30% for global bonds, IAUM at 10% for gold. For the gold sleeve to drift 5 points, it has to fall by half or grow to 15% of everything you own. By the time a band that wide triggers, the position has stopped resembling the one you sized.

Larry Swedroe's 5/25 rule applies whichever band is tighter: 5 percentage points, or 25% of the sleeve's own target weight. The 10% gold position gets a 7.5% to 12.5% band. VT keeps its 55% to 65% one, because 25% of 60 is bigger than 5.

Twenty percent of the portfolio is where the two bands cross, since 25% of 20 is 5. Above that line a sleeve keeps the plain 5 points; below it, the proportional band takes over. Split the equity side into VTI for US stocks and VXUS for everything outside it, and a 40/20 split leaves both on 5 points, while a 45/15 one puts VXUS on a 3.75-point band. A core-satellite portfolio writes its satellite cap down for the same reason: a small sleeve needs a rule of its own.

Rebalance with new money before you sell

Selling is the expensive way to do it, so work down the list before you get there.

New money triggers no tax. Pointing each contribution at whichever sleeve sits furthest below target closes the gap without a sale, and for anyone still adding to the portfolio it does most of the work on its own. That mechanic gets its own treatment in dollar-cost averaging with ETFs.

If part of the portfolio sits in an account that shelters gains from tax, do the selling there. Vanguard modelled a portfolio split evenly between sheltered and taxable accounts and found that rebalancing on the sheltered side first improved after-tax returns by 44 basis points a year, with no increase in volatility. Which fund belongs in which account is a related decision, covered in asset location.

When you do have to sell in a taxable account, two things soften it. If your tax rules let you choose which shares go (US brokers call it specific-lot identification; the UK, France and Canada pool your cost basis instead), sell the highest-cost ones, since they carry the smallest gain. And you don't have to go all the way back to target: moving from 68% to 63% captures most of the risk correction for a fraction of the tax bill. The tax cost of selling a winner is usually the largest number in the whole exercise.

Beacon's portfolio tool holds the target weights for you and reads back how far off you are. Fill in the shares you hold, and it points your next contribution at the sleeve that's furthest behind.

Write the rule down before you need it

The moment a band gets crossed is the moment acting on it feels worst.

The crisis bottomed in March 2009. A 60/40 rule was telling you to buy equities exactly when that felt least survivable. Vanguard's read of that decade is that many investors were bearish enough by then to lack the confidence to rebalance toward stocks without a consistent rules-based approach. That was when the drifted portfolio was costing them the most. Decide the rule in a calm month, because the calm version of you is the one you want making that call.