Allocation Drift: The Silent, Measurable Risk You Can't See by Hand

Allocation drift is the widening gap between the weights you intended for each part of your portfolio and the weights you actually hold. You set out to keep a strategy at 25% of the book; a good run pushes it to 32% while you do nothing, and now a third of your risk sits somewhere you never chose. The danger isn't that drift is hidden — it's that it's invisible to the eye but perfectly visible to arithmetic. It's a number you can compute from your own account every week: current weight minus target weight, per strategy, in percentage points. This is what that number is, how to calculate it, and the rule for when a breach should actually make you act.

I run a leveraged, income-oriented book myself, and drift is the failure mode I respect most — precisely because it never feels like a mistake. Nobody decides to become overconcentrated. You just keep winning in one place and skip the boring step of measuring, and the portfolio quietly rewrites its own risk profile. The fix isn't willpower. It's a weekly number and a rule.

What is allocation drift?

Allocation drift is the change in your portfolio's mix that happens on its own, because the parts grow at different rates. You choose target weights — how much of the book each strategy should carry — and then the market moves them. The winners swell, the laggards shrink, and the real portfolio slowly diverges from the one you designed. Drift is simply the size of that divergence, measured strategy by strategy. None of it is even measurable until the book is sorted into strategies with distinct jobs rather than one long list of names — the same reason a pile of good stocks isn't a portfolio.

The reason it matters is that weight is risk. When a strategy grows from a quarter of your book to a third, you didn't just make money — you increased how much a bad run in that one strategy can cost you. Your risk went up without a single decision, and it went up quietest in exactly the strategies that were doing best. In the Incomestead Stack — our name for the whole target allocation across strategy Layers — drift is what pulls a Layer away from the weight you assigned it.

Drift is the only risk that grows while you're doing nothing — and feels like success the whole time it's happening.

Why can't you see allocation drift by hand?

Because your brokerage screen is built to show you the wrong thing. It shows positions and dollar values — this many shares, worth this much, up this percent. What it never shows is each strategy's share of the whole book measured against the target you set for it. That comparison is the entire signal, and no default account view computes it.

So the drift hides in plain sight. A position doubling looks like good news, not a risk event, because nothing on the screen frames it as "this Layer is now eight points over its target." To see drift you have to leave the broker's view and do four steps by hand every week: write down each Layer's target, group every holding into its Layer, total each group as a percent of the book, and subtract the target. Do it once and you'll understand why almost nobody sustains it — it's fiddly arithmetic that quietly decays the first week you're busy. That decay is the single point of failure: the measurement depends entirely on you remembering to do it, perfectly, forever. That's exactly the job a governed process exists to take off your shoulders.

How do you measure allocation drift?

The core calculation is deliberately simple — it has to be, or it won't survive contact with a busy week:

Two things turn that formula into a usable instrument. First, the tolerance band: a width around the target inside which you deliberately do nothing. Real portfolios are never exactly on target, and chasing every fractional wobble just churns costs. A band — say ±5 percentage points — says "anything this close is fine; only flag what's genuinely out of line." Vanguard's rebalancing research made the honest version of this point: there's no single magic rebalancing frequency or threshold, and rebalancing is fundamentally about keeping risk aligned with your target, not squeezing out extra return (Vanguard, "Getting Back on Track: A Guide to Smart Rebalancing"). Bands are a risk-control tool, not an alpha engine — set yours for how much divergence you can live with, not to time anything.

Second, drift only means something measured on a cadence. A single reading tells you where you are; a series of readings tells you whether you're moving, how fast, and whether a breach is a blip or a trend. That same Vanguard research found no material difference in long-run risk or return across rebalancing frequencies from monthly to annual — what matters is having a rule and reading against it on a set rhythm, not which calendar day you pick (Vanguard, "Getting Back on Track"). The number is only as good as the rhythm you read it on.

What drift looks like over six weeks

Here's the payoff — and why "you can't see it by hand" is literally true. Take a $2,400,000 book with a Long-Horizon Growth Layer targeted at 25% and a tolerance band of ±5 points (so anything from 20% to 30% is in-band). It has a strong run. Watch what the weekly drift number does while nothing on the brokerage screen ever flashes red:

In dollar terms, that Layer went from about $600,000 to roughly $792,000 (as the other Layers slipped) of the book over six weeks — a $190,000-odd swing in how much of your future rides on one strategy. Nothing broke. No alarm sounded. Every week looked like a good week. The only thing that caught it was a number computed the same way each time and compared to a line drawn in advance. That is the whole argument for measuring drift: the risk built up in the weeks that felt best.

By the week it finally crossed the line, the position had been drifting for over a month. The breach wasn't the event. The month was.

When should drift actually trigger a rebalance?

Not on the first breach. A single weekly reading can cross a band on noise — a volatile week, a dividend, an options assignment — and reverse the next week. If you rebalance on every first flag, you'll churn costs and trade against yourself. So the rule is the two-week persistence rule: a breach must show up in two consecutive weekly snapshots before it earns an action. In the table above, Week 5 flags the breach and Week 6 confirms it — that's when a governed book proposes a move, not before.

And the move itself is bounded, not dramatic. The disciplined response to an overweight Layer is to trim it back toward the midpoint between its target and the band edge — here, halfway between the 25% target and the 30% upper bound, so about 27.5% — not all the way to target (which over-trades) and not to zero (which abandons a working strategy). That gives you a specific, unemotional instruction: sell down roughly enough of the Layer to bring it to ~27.5% of the book. A recommendation, to be clear — you decide whether and how to execute it.

One precedence note, because it's the rule that outranks this one: if your margin is also out of line in the same week, margin comes first. Leverage is the existential risk; allocation is the important-but-survivable one. Fix the margin breach, then address the drift.

How allocation drift and margin fit into one score

Allocation drift and margin utilization are the two things in a leveraged book that move while you're not looking — and they're governed the same way: measure weekly, compare to a line set in advance, act only on a persisted breach. Both have a browser tool of their own in the free tools, if you would rather not do the arithmetic by hand. Margin just uses zones instead of a band — Clear below 30% utilization, Harvest 30% to under 35%, Freeze 35% to under 40%, Forced 40% and above — and the same two-week persistence logic before a zone change drives action. If you want the leverage side in full, that's the weekly margin management piece.

Both roll up into one number we call the Governance Score — a single read on how well you're running the book, blending allocation drift, margin zone, and whether you're actually checking on cadence. Drift is the allocation component. A book that's quietly seven points over on one Layer can't be governing well, however good the individual picks look — and the Score is built to say so, in a way a pile of green P&L lines never will. That's the shift the whole system is for: from asking "am I diversified enough?" once in a worried moment to reading a governed number every week.

Measure your own drift this week

You can compute your real drift in about fifteen minutes, no software, using your own account:

The first time you run it, the surprise is usually not the drift itself but how long it had been building unmeasured. That's not a failure — it's the moment the instrument panel turns on. Almost every operator I've compared notes with, myself included, has found one Layer meaningfully off-target the first week they actually measured, and none of them were careless. They just didn't have the number. Once you do, the anxious question — am I quietly overexposed? — becomes a weekly line item you either clear or act on, instead of a worry you carry.

"I'll rebalance when it feels off" is a plan to notice too late. A band and a cadence notice for you.

None of this is a recommendation to rebalance on any particular schedule or to any particular target — the ±5 band and the 25% weight here are illustrations, and the right numbers for your book are yours to set. Rebalancing also has real costs and, in taxable accounts, tax consequences that are entirely outside this page and belong with your CPA. The point is narrower and it's the whole brand: drift is measurable, so stop guessing at it. Turn it into a governed number you read every week.

Frequently asked questions

What is allocation drift?

Allocation drift is the gap that opens between your intended target weights and the weights you actually hold, because different parts of the portfolio grow at different rates. Measured per strategy, it's current weight minus target weight, in percentage points. It raises your risk without any decision on your part — a winning strategy you never trim can quietly become a much larger share of the book than you chose.

How do you measure allocation drift?

Total your whole portfolio's value, group holdings by strategy, divide each group by the total to get its current weight, then subtract that strategy's target weight. The result — in percentage points, with a sign — is its drift. Compare each figure to a tolerance band (for example ±5 points); anything beyond the band is a breach worth attention.

What is a good rebalancing threshold or tolerance band?

There's no single correct number, and this isn't personalized advice. Research on rebalancing has found no universally optimal frequency or threshold — bands are a way to control risk and avoid needless trading, not to boost returns. Many investors use something in the range of a few percentage points around each target; the right width depends on how much divergence you're willing to hold and your own costs. Set it deliberately in advance rather than deciding in the moment.

How often should I check for drift?

Drift only means something read on a cadence, because a series of readings — not a single one — tells you whether a breach is a blip or a trend. A weekly snapshot is enough to catch drift early while filtering noise, especially when paired with a persistence rule: only act when a breach shows up in two consecutive readings.

When should allocation drift actually trigger a rebalance?

When a breach persists. A single reading can cross a band on noise and reverse the next week, so a disciplined approach waits for the breach to appear in two consecutive weekly snapshots before acting — and then trims the overweight strategy back toward the midpoint between its target and the band edge, not all the way to target. If margin is also out of line the same week, address margin first.

- This is education, not personalized financial, tax, or legal advice. I write as a practitioner who runs a leveraged, income-oriented book, not as a licensed adviser. Rebalancing has transaction costs and, in taxable accounts, tax consequences that are outside the scope of this article; consult a qualified professional before acting. The weights, bands, and figures here are illustrative, not a recommendation for your money.

Incomestead recommends. You decide.