Portfolio drift: why the mix you chose stops being the mix you hold

The question almost nobody asks

Most people who invest make one big decision and then stop thinking about it. They pick a split: how much in shares, how much in bonds, maybe a slice of cash. That split has a name — asset allocation, meaning simply the proportions your money is divided into [11]. It is not a small choice. One widely cited finding holds that the split explains more than 90% of the variation in how a portfolio behaves over time [7].

Then the years pass. Nothing is bought. Nothing is sold. And the split quietly changes anyway.

Here is the question worth asking: what mix are you actually holding right now? Not the one you picked. The one you have.

What actually happens

The mechanism is arithmetic, not psychology. If two things you own grow at different speeds, the faster one becomes a bigger share of the total. That is all. No decision is required, and no decision is recorded.

Morningstar puts a number on the last decade. US shares compounded at more than 14% a year on average over the ten years to late 2025, while bonds barely moved. A portfolio that began at 60% shares and 40% bonds ten years ago, left completely alone, would now hold more than 80% shares [3]. Its own strategists put it plainly in a subscriber session: a 60/40 split, left unchecked for a decade, would likely shift to 80% shares [6].

The same thing happens on shorter clocks. Investopedia’s worked example takes a portfolio built at 80% shares and 20% bonds; after a single strong year it is closer to 85/15 [12]. Kiplinger’s version starts at 60/40 and lands near 65/35 [14]. The direction is always the same: toward whatever has been winning.

Fund managers have a name for their own version of this. It is called style drift, and Investopedia notes it “can result naturally from capital appreciation in one asset relative to others” — no manager decision needed [18]. The problem is considered serious enough that the securities regulator requires a fund to keep 80% of its assets in what its name implies [18].

The numbers, worked through

Round numbers, in the shape of the last decade. Start with $100: $60 in shares, $40 in bonds.

Say shares grow three and a half times over ten years. Bonds gain 15%.

• Shares: $60 becomes $210.
• Bonds: $40 becomes $46.
• Total: $256.

Shares are now $210 of $256. That is 82%.

You started at 60%. You are holding 82%. The mix moved 22 percentage points, and you made zero decisions. These figures are illustrative, not a forecast — but the direction is not optional. Whenever one holding grows faster than another, this happens.

It happens a second time, inside the fund

Here is the layer most people miss. Even someone who owns nothing but a single broad index fund has drifted — because the index itself drifted.

The largest US shares have dramatically outrun the broad market since the pandemic, and as their value rises they take up more of any index that holds them. The top 10 companies now make up roughly 40% of the S&P 500’s total value, which Morningstar’s chief multi-asset strategist calls a “highly unusual” level against history [20]. He adds that many investors are probably unaware how concentrated their holdings have become [20].

Set that against a common rule of thumb. Morningstar’s director of personal finance suggests 5% to 10% in any single holding “gets to be a lot” [10]. The top ten names of the main US index are collectively four times that. People who also own growth or sector funds often own the same handful of shares several times over without noticing [20].

What stepping in costs, and what it buys

Bringing the mix back to target is called rebalancing: selling some of what grew and buying what lagged [1] [5]. The honest question is what it actually achieves, and the answer is narrower than the sales pitch.

Morningstar’s Christine Benz is direct about it. Researchers have looked at whether rebalancing boosts returns, she says, but “risk reduction is really the main reason to consider” it [15]. Morningstar’s own guidance says the same thing more bluntly: rebalancing “doesn’t necessarily improve your portfolio’s returns, especially if it means selling asset classes that continue to perform well” [3].

That is the trade. You give up some of the winner’s future run in exchange for a mix that behaves the way you expected. Whether that is worth it depends entirely on what the money is for. Benz notes it matters far more the closer someone gets to needing the money, and far less for someone in their twenties [15].

The costs are real but smaller than they were. Inside a tax-sheltered retirement account, moving between holdings triggers no tax at all — tax comes only on withdrawal [21]. In an ordinary taxable account, selling a winner usually realises a gain the tax authority wants a share of [21]. Trading has “gotten a lot less expensive,” Investopedia notes, but is not free [12].

Where people go wrong

Treating the calendar as the signal. Quarter-end rebalancing “relies on arbitrary calendar events which may not coincide with market movements,” Investopedia points out — and the widespread belief that big institutions always rebalance at quarter-end has never actually been evidenced [5]. The CFA Institute recognises two approaches: fixed calendar dates, or bands that trigger when the mix strays past a set distance. Advisers have leaned toward bands in recent years [12]. Benz suggests 5 or 10 percentage points from target as a reasonable trigger [15].

Assuming automation means untouched. Target-date funds do the resetting for you, typically once a year, along a planned path that shifts toward bonds as the date nears [2] [16]. They now hold $4.8 trillion in the US, up 20.3% in a year [17]. They are not free: the fund-of-funds structure adds a layer of fees [16]. The average cost has fallen to 27 hundredths of a percent, roughly half its level a decade ago [17].

Believing the mix is the only thing that drifts. Concentration in a single employer’s shares is its own version, and one a simple rebalance does not fix [14] [13] [8]. Benz warns that a paycheque and a share holding from the same company are one bet, not two [10].

What is genuinely uncertain

The behavioural evidence points in two directions at once, and it is worth being honest about that. The documented disposition effect says investors tend to sell winners too early and cling to losers too long [19] [9]. Yet drift data says untouched portfolios end up stuffed with winners. Both are real — one describes trades people make, the other describes the holdings they never touch.

There is also a live argument about whether the classic 60/40 still does its job. Shares and bonds have moved together more often lately: they fell together in about 14% of months over 25 years, but 28% over the last five [4]. Morningstar’s counter is that correlation says nothing about the size of moves, and that the split was never meant to cancel every fall, only to soften it [4].

And the deepest uncertainty is not financial at all. A portfolio that drifted to 82% shares is only a problem if 82% is more risk than the owner can carry. Sometimes it is not — a longer horizon or a bigger cushion genuinely changes what a person can absorb [15] [10]. The issue is rarely that the number changed. It is that nobody chose it.

Morningstar’s Mind the Gap study puts a price on the general problem of investors and their own timing. The average dollar in US funds earned 8.7% a year over the ten years to the end of 2025. The funds themselves returned 9.9%. That gap of 1.2 percentage points is about 12% of the return [22]. The funds people did least to, including the ones that rebalance automatically, tended to close that gap best [22].

The one thing to carry

Do nothing and your mix will still change. It will change toward whatever has been winning, and the change will be largest right after the longest run. The mix you are holding is a fact you can check today. It is not the same thing as the mix you picked.