> ## Content Index > Fetch the complete content index at: https://www.incomestead.com/llms.txt > Use this file to discover other available public pages before exploring further. # How Often Should I Rebalance My Portfolio? The Frequency Isn't the Rule - URL: https://www.incomestead.com/blog/how-often-should-i-rebalance-my-portfolio/ - Published: 2026-09-03T14:42:37.000Z - Updated: 2026-09-11T12:16:10.000Z - Description: Every answer to this question is a frequency, and a frequency alone is a calendar reminder you will ignore. The two-instrument version instead. - Author: Stefano Starkel - Tags: The Stack, #stage-2 **How often should you rebalance your portfolio?** Review it weekly; act only when a holding breaches its band and is still breached the following week. Every answer you will find is a frequency — annually, quarterly, on your birthday — and a frequency on its own is a calendar reminder you will eventually ignore, because nothing in it tells you whether *today* is a day to act. What survives contact with a real book is two instruments rather than one: a **review cadence**, how often you look, and an **action threshold**, what actually makes you trade. Confusing them is how you end up over-trading and drifting at the same time. I run a leveraged, income-oriented book myself, and rebalancing is where I learned the frequency question is a decoy. My cadence never changes — fifteen minutes, same slot, every week. What changes is whether anything crossed a line. Most weeks nothing has, and the correct action is none. TL;DR - **A frequency is not a rule.** "Once a year" tells you when to look and nothing about when to act. - **Two instruments, not one.** A *cadence* (fixed, weekly, cheap) and a *threshold* (a band around each target). The cadence finds the breach; the threshold decides it. - **Bands are proportional, not fixed points** — 30% either side of the target — so a big Layer gets a wide band and a small one a narrow band. - **Don't act on one reading.** A breach must persist two consecutive weekly snapshots — the **two-week persistence rule** — before it drives a trade. - **On a leveraged book, *how* you rebalance is a margin decision.** One breach, corrected to the same weight three different ways, lands in three different margin zones: 14.85%, 30.23% or 40.29%. On this page - [How often should you rebalance a portfolio?](#how-often-should-you-rebalance-a-portfolio) - [Why is an annual rebalance an answer to the wrong question?](#why-is-an-annual-rebalance-an-answer-to-the-wrong-question) - [The two instruments: a cadence you keep, and a threshold that fires](#the-two-instruments-a-cadence-you-keep-and-a-threshold-that-fires) - [What does a real breach look like on a seven-figure book?](#what-does-a-real-breach-look-like-on-a-seven-figure-book) - [What changes when the book carries a margin loan?](#what-changes-when-the-book-carries-a-margin-loan) - [Write your rebalancing rule this week](#write-your-rebalancing-rule-this-week) - [Frequently asked questions](#frequently-asked-questions) ## How often should you rebalance a portfolio? Review it **weekly**. Rebalance it **when something breaches its band and stays breached** — which happens when it happens, not on a schedule you could predict in January. How often that is depends entirely on how wide your bands are and how the year behaves. That is one answer to two questions, and the usual advice feels unsatisfying because it only answers the first. "Rebalance annually" is a review cadence wearing a trade's clothing: it tells you to open the account on a date, not what you are looking for or what would make you act. The distinction bites in both directions. In a quiet year an annual rebalance makes you trade for no reason — you rebalance because it is January, not because anything moved. In a violent year it makes you carry a breach for months, because the position that ran away from you in March is corrected the following January, long after the risk you never chose has been taken. > A calendar tells you when to look. On its own it cannot tell you when to act. ## Why is an annual rebalance an answer to the wrong question? Because the interval was never the risky part. The risky part is the gap between the weights you chose and the weights you actually hold, and that gap does not open on an annual schedule. It opens whenever one part of the book does well for long enough that nobody notices it has become a different portfolio. That gap is [allocation drift](https://www.incomestead.com/blog/allocation-drift/), and it is a number rather than a feeling: current weight minus target weight, per Layer, in percentage points. That post owns both the measurement and the question of which threshold should trigger action. This one is about the other half — the cadence, and what the frequency choice actually costs you. Annual rebalancing is not wrong, and I want to be straight about where it is genuinely the right call. If your book is a handful of broad funds with no leverage, if you hold it somewhere every trade carries a cost you would rather not incur, and if you know from experience that you will actually do the annual review rather than skip it, then a once-a-year discipline you keep beats a weekly one you abandon in month three. **The best cadence is the one you will still be running in five years.** What I am arguing against is not the annual date. It is the belief that the date is the rule. ## The two instruments: a cadence you keep, and a threshold that fires Separate them and both get easier to hold. The **cadence** is how often you look: fixed, frequent, boring — mine is weekly, and [the weeks you skip are the dangerous ones](https://www.incomestead.com/blog/the-weeks-you-skip/). A weekly reading costs almost nothing precisely because it usually ends in no action. Its job is not to generate trades; it is to make sure that when something moves, you find out in days rather than quarters. The **threshold** is what makes you act. In our framework it is a **proportional** band rather than a fixed number of points — 30% either side of each target, so a large Layer gets a wide band and a small one a narrow band. [The drift post](https://www.incomestead.com/blog/allocation-drift/) derives why proportional beats fixed and where exactly the line sits; here it is enough to know the band scales with the target, and that you set it once rather than judging it in the moment. Two things then stop the threshold becoming a hair trigger. First, when a breach is corrected you do not go back to target — you go to the **midpoint of the target and the edge that was breached**, which leaves the position room to move without immediately re-breaching. Second, a breach must **persist across two consecutive weekly snapshots** before it drives a trade. That is the two-week persistence rule, and it exists because one ugly print is not a decision. Both rules are in the [published rule set](https://www.incomestead.com/governance-specification/). ## What does a real breach look like on a seven-figure book? Here is the shape, with round numbers you can check against your own account. The target weight below is an *illustrative* policy — I picked it to make the arithmetic legible. The band rule around it is not illustrative; it is the actual rule. Say the book holds **$1,720,000** of securities at market. One Layer — call it Long-Horizon Growth — has a target weight of **45.00%**. Proportional bands put its ceiling at 45.00 × 1.30 = **58.50%** and its floor at 45.00 × 0.70 = **31.50%**. After a strong run that Layer is now worth **$1,040,000**, which is **60.47%** of the book. It is **1.97** percentage points above its ceiling. That is a breach. Two things this example holds still that the real world will not. Prices move between your two readings, so the persistence rule means you will usually act on slightly different numbers than the ones that flagged the breach — that is a reason to size the trade at the second reading, not the first. And every figure below ignores commission, spread and the interest accruing on the loan. None of that changes the shape of the argument; all of it changes the last two decimal places. Notice what the weekly cadence just did and did not do. It *found* the breach. It did not authorise a trade — under the two-week persistence rule, the same reading has to come back next week before anything happens. If the Layer subsides below 58.50% in the meantime, the right action was none, and a calendar-driven rebalance would have traded anyway. Say it persists. The correction does not take the Layer back to 45.00%. It takes it to the midpoint of the target and the breached edge — (45.00 + 58.50) ÷ 2 = **51.75%**, or **$890,100** of a $1,720,000 book. So the instruction is to move **$149,900** out of that Layer. In a taxable account that is a sale of a winner, with the tax consequences that implies — out of scope here, but not out of mind. Not to target. To the midpoint, deliberately, so a normal week does not put you straight back over the line. > The cadence finds it. The threshold decides it. Persistence confirms it. No one of those three is a rebalancing rule on its own. ## What changes when the book carries a margin loan? This is the part general rebalancing guidance tends to skip, and it is the part that can hurt you. Most of what is written on this question assumes an unlevered portfolio, where a rebalance is a neutral swap: sell some of this, buy some of that, nothing else moves. **On a book carrying a margin loan that assumption is false, because the route you take to the same destination changes your leverage.** Add a **$520,000** margin loan to the same book, held in a taxable margin account — this section does not apply to an IRA, where you cannot borrow like this at all. Margin utilization, as we define it, is the loan divided by the gross value of the securities — the market value of everything you hold, before the loan is netted, with no cash term, because when you carry a margin loan the loan *is* your negative cash balance. That is **$520,000** ÷ **$1,720,000** \= **30.23%**, which sits in Harvest. Be careful carrying that number between sources: "margin utilization" is a real term elsewhere, but brokers who publish one may compute it differently — some against margin available, some net of the loan — so it is worth checking which denominator you are reading before you compare. ClearHarvestFreezeForced | Zone | Utilization | What it means | | ------- | ---------------- | ---------------------------------------------------------------------------------- | | Clear | under 30% | Clear. Room to absorb a drawdown without acting. Where the disciplined book lives. | | Harvest | 30% to under 35% | Elevated. Direct new income to paying the loan down; stop adding leverage. | | Freeze | 35% to under 40% | No new borrowing, no new risk. Actively reduce the loan. | | Forced | 40% and above | Existential. Reduce now, on your terms — before the broker does it on theirs. | House limits, not a regulatory standard — these bands are deliberately tighter than any broker or bank requirement. Now correct the same breach three ways. All three end with that Layer at exactly **51.75%** — the same breached weight, brought back by the same rule. (The *rest* of the book differs between them, and so does the size of the sale: reaching the midpoint by selling also shrinks the book you are measuring against, so route B has to sell **$310,674** where route A sells **$149,900**.) Watch the margin reading. | Route | What you do | Loan | Gross securities | Utilization | Zone | | ----- | --------------------------------------------------------------------------------- | -------- | ---------------- | ----------- | ----------------------- | | **A** | Sell $149,900 of the overweight Layer, buy the underweight ones with the proceeds | $520,000 | $1,720,000 | **30.23%** | Harvest — unchanged | | **B** | Sell $310,674 and stop — the proceeds pay down the debit | $209,326 | $1,409,326 | **14.85%** | Clear — you deleveraged | | **C** | Sell nothing — buy $289,662 of the underweight Layers on margin instead | $809,662 | $2,009,662 | **40.29%** | Forced | > Same breach. Same corrected weight. Same rule. Three zones — because nobody wrote down which route to take. Three routes, one corrected weight, and a spread from **14.85%** to **40.29%** — from the calmest zone to the most dangerous one, off a single breach. Route C is the one to sit with. It is not exotic or reckless-looking; it is what "I'll rebalance by topping up the laggards instead of selling my winner" turns into when the account has margin available. It fixes the allocation and creates a leverage problem in the same trade. And Forced is not just a worse-sounding word. It is the one zone that acts on its *first* reading — the two-week persistence rule that protects you everywhere else does not apply there, because at that level the risk is existential rather than annoying. It is also why our framework resolves conflicts **margin first, allocation second**: if fixing your weights breaks your leverage, you have not fixed anything. Route C would also cap the whole [Governance Score](https://www.incomestead.com/blog/governance-score/) at 34 — the top of its worst band — so a book that just did the textbook-correct thing to its allocation reads *Exposed*. These four bands are ours, not a regulator's, and they are deliberately far tighter than anything that triggers mechanically — so it is worth showing that gap rather than asserting it. Under this convention your equity is what is left over: at **40.29%** utilization you still hold **59.71%** equity. FINRA's maintenance floor is that *"a customer's equity in a margin account that holds margin stocks only must not fall below 25 percent of the current market value of the long securities"* ([FINRA, margin calls](https://www.finra.org/investors/insights/margin-calls?ref=incomestead.com)) — note the scope, which is a plain margin-stock account, not an options or multi-strategy book. Twenty-five percent equity is **75.00%** utilization. Our Forced band sits about thirty-five points below it. That gap is real, but it is not a promise, and the same FINRA page says why. Firms *"may also set margin requirements of their own—often called 'house' requirements—that can be higher than the margin requirements under Reg T"*, and *"firms may increase the house requirements at any time and aren't required to provide you with advanced written notice."* A common house floor of 30% equity is **70.00%** utilization — still clear of our bands, but it can move, in the week you would least like it to. Two more sentences from that page: *"Firms don't have to issue a margin call before selling securities in your margin account to meet a margin call"*, and *"firms don't have to let you choose which securities or assets are sold to meet a margin call."* Route C is not dangerous because it puts you near a call. It is dangerous because it quietly gives away the thing this whole post is about — choosing. ## Write your rebalancing rule this week Not a frequency. A rule. Four lines, written down — because a rule you hold only in your head gets renegotiated by the market in exactly the week it matters. That document has a name here: the **Investment Charter**, and there is a [guided version](https://www.incomestead.com/blog/personal-investment-policy-statement/#tool-investment-charter) if you would rather not start from a blank page. The four-line rebalancing rule 1. **Cadence.** "I read the book every **, for fifteen minutes." Pick the day. Fixed and frequent beats clever. 2. **Threshold.** "A Layer is breached when it leaves its band." Write each target and its two bounds — target × 0.70 and target × 1.30 — so you never have to judge in the moment. 3. **Persistence.** "A breach acts only if it is still there next week." One reading is information; two is a decision. 4. **Route.** "When I correct a breach I will **, and I will not fund a rebalance by borrowing." This is the line that would have stopped Route C, and it is the one almost nobody writes. > A rebalance you fund by borrowing is not a rebalance. It is a leverage decision wearing a rebalance's clothes. Line four is the one worth arguing with yourself about. If your book carries no leverage it costs nothing to write. If it does, it is the whole difference between a rebalance and a leverage decision you did not know you were making. Do this next Before you pick a cadence, find out whether anything is actually breached right now. The **Allocation Reality Check** takes your holdings and your targets and returns your real drift — the number this whole post is a rule for. [Run the Allocation Reality Check →](https://www.incomestead.com/blog/allocation-drift/#tool-allocation-reality-check) ## Frequently asked questions ### How often should I rebalance my portfolio? Review it weekly; rebalance only when a holding breaches its tolerance band and is still breached at the next weekly reading. The frequency you hear quoted — annually, quarterly — is a review cadence, not a rebalancing rule. Separating the two means you stop trading on dates when nothing has moved, and stop carrying a breach for months when something has. ### Is it better to rebalance on a calendar or on a threshold? You need both, doing different jobs. The calendar sets how often you look, which should be frequent and fixed. The threshold sets what makes you act, which should be a band around each target weight — in our framework a proportional one, running from the target multiplied by 0.70 up to the target multiplied by 1.30\. A calendar with no threshold trades for no reason; a threshold with no calendar never gets checked. ### Does rebalancing affect margin? It depends entirely on how you do it, which is why the route belongs in your written rule. Margin utilization is your loan divided by the gross market value of your securities, before the loan is netted. Selling an overweight holding and buying underweight ones with the proceeds leaves both numbers untouched. Selling and holding the proceeds pays down the loan and lowers your utilization. But funding the correction by *buying* the underweight holdings on margin raises the loan and the market value together, and on the book worked through above that single choice moves the same book from 30.23% utilization to 40.29% — from Harvest, where you simply stop adding leverage, into Forced, the existential zone that acts on its first reading rather than waiting the usual two weeks. ### What if I only rebalance once a year? That can be entirely reasonable, and it is better than a weekly discipline you abandon. If your book is unlevered, simple, and held somewhere trading has a real cost, an annual review you actually perform is a defensible operating model. Write the threshold down anyway, so that when you do sit down you are comparing your weights against a rule rather than against your memory of what you meant to hold. This is education, not personalized advice. I run a leveraged, income-oriented book myself and write from that experience; I am not a licensed adviser. Margin borrowing carries real downside: your broker can sell your holdings to meet a shortfall, and is not required to contact you first. The target weight in the worked example is illustrative — the band, midpoint and persistence rules around it are the ones I actually run. Nothing here is a recommendation about how much you should borrow or hold, and tax consequences of rebalancing are outside this post's scope — take those to a qualified professional. Incomestead recommends. You decide.