Most portfolio drift problems are not solved by knowing how to rebalance — they are solved by deciding when and under what conditions to rebalance before the moment arrives. Without a pre-committed policy, that decision gets made emotionally: either too often (after every leg down, when anxiety peaks) or never (because selling a winner feels wrong when prices are rising). Both patterns have the same root cause — the absence of a written rule.

This article is the action layer of a two-part pair. The companion piece, Portfolio Review Cadence, covers how to observe your portfolio systematically over time — the scheduled look. This article is about what you do when the look reveals drift: drafting a pre-committed policy that removes the in-the-moment decision and replaces it with an execution. By the end, you will be able to write a rebalancing policy, distinguish two major policy types, and reason explicitly about the three-way trade-off every policy makes.

What a Rebalancing Policy Is — and Why Pre-Commitment Is the Point

A rebalancing policy is a written rule that specifies, in advance, when and how you will return a portfolio's weights toward its target allocation. Target weights are the intended share of the portfolio each position should hold — the allocation you chose when you were calm, deliberate, and not watching prices move.

Left alone, a portfolio's weights drift. Winners grow into a larger share; losers shrink. That is not a malfunction — it is arithmetic. But drift changes the risk profile you originally set. A portfolio you designed to hold equal shares of four positions gradually becomes a concentrated bet on whatever has performed best. The risk you are carrying is no longer the risk you intended to carry.

Rebalancing is the act of returning toward target weights. The policy is the part that matters most, because it governs when that act happens. Without one, rebalancing becomes a reactive response to emotional state: you sell winners when a pullback makes you nervous, or you never rebalance because the portfolio "looks fine." A pre-committed policy removes that choice from the hot moment and places it in the hands of the version of you who is calm, clear, and not watching prices move in real time.

This connects directly to the broader principle covered in Pre-Commitment and If-Then Rules: decisions made in advance under cool conditions consistently outperform decisions remade under pressure, even by people who know their own tendency to renegotiate. Writing a rebalancing policy is an application of exactly that principle to portfolio structure.

Two Policy Types: Calendar and Threshold

There are two standard design approaches to a rebalancing policy. They are not mutually exclusive — some practitioners combine them — but understanding each in isolation clarifies what they are optimizing for.

Calendar (time-based) rebalancing means you act on a fixed schedule: once per quarter, once per year, or at whatever interval you have pre-committed to. On that date, you check the weights and return them toward target, regardless of how much drift has occurred. The calendar does not care about the market environment.

The advantage is simplicity. You do not need to monitor the portfolio between scheduled dates, there is no ambiguity about whether a trigger has been reached, and the rule is easy to follow because it is time-driven rather than magnitude-driven. The trade-off is that a calendar policy may rebalance when drift is minimal (wasting transaction cost on an unnecessary action) and may let significant drift accumulate between reviews if prices move sharply close after a scheduled rebalance.

Threshold (tolerance-band) rebalancing means you act when a weight drifts beyond a defined band around its target. A 25% target position with a band of ±10 percentage points would trigger rebalancing only if that position's share reaches 35% or falls to 15%. Until that threshold is crossed, the policy says: do nothing.

The advantage is proportionality. Rebalancing is triggered by actual drift, not by calendar proximity. You do not pay transaction costs when the portfolio has barely moved. The trade-off is that this approach requires more frequent monitoring — you need to check whether a threshold has been crossed — and choosing the band width involves a judgment call. A very tight band triggers frequently (high cost, high drift control). A very wide band triggers rarely (low cost, poor drift control).

A combined approach might set both a calendar review and a threshold: check on a schedule, but act only if drift exceeds the band. That hybrid reduces unnecessary trades while ensuring the review actually happens.

All band widths used as examples throughout this article, including any specific numbers, are illustrative only. They are not recommendations for any particular portfolio or investor.

The Three-Way Trade-Off: Drift Control, Cost, and Effort

Every rebalancing policy is balancing three competing interests simultaneously. Understanding this trade-off is what makes the policy decision real rather than mechanical.

Drift control is how closely the portfolio tracks its target allocation over time. Tight drift control means rebalancing early and often; the portfolio's actual weights rarely stray far from intended weights. Loose drift control means accepting more deviation — which means the risk profile you are carrying may diverge meaningfully from the risk profile you intended to carry.

Cost and turnover are the friction created by each rebalancing event. Every trade has a transaction cost of some kind — explicit (commissions, spreads) or implicit (taxes triggered on realized gains, time spent executing). Rebalancing frequently keeps drift tight but multiplies these costs. Rebalancing rarely keeps costs low but allows more drift.

Effort is the monitoring and decision burden. A threshold policy requires checking regularly whether a trigger has been crossed. A calendar policy requires less ongoing attention but demands disciplined execution on schedule.

Vanguard research (Jaconetti, Kinniry & Zilbering) makes the relationship between these factors unusually clear. The researchers state explicitly: "The primary goal of a rebalancing strategy is to minimize risk relative to a target asset allocation, rather than to maximize returns." And on the question of which policy is best: "there is no optimal frequency or threshold for rebalancing, since risk-adjusted returns do not differ meaningfully from one rebalancing strategy to another."

That second finding is important enough to sit with. Whether you rebalance monthly, quarterly, or annually, the risk-adjusted outcomes across those approaches are similar. What does differ significantly is the number of rebalancing events — and therefore the costs — as frequency increases. A monthly policy will produce many more transactions than an annual one without producing meaningfully better risk-adjusted outcomes. The Vanguard research concludes that cost minimization, alongside risk control, is a primary practical consideration when choosing frequency: more frequent rebalancing does not produce meaningfully better risk-adjusted outcomes, so the additional transaction cost it generates is hard to justify.

No rebalancing policy improves returns. That is not what rebalancing does. It manages risk by maintaining the risk profile you chose. Anyone who tells you a particular rebalancing approach generates alpha is making a claim the evidence does not support.

The practical design implication: choose the least frequent rebalancing schedule that still keeps your drift within a tolerance you can actually accept. More frequent rebalancing does not pay for itself in better outcomes — it only pays for itself if you genuinely cannot tolerate the drift a less frequent policy produces.

Position sizing principles connect here directly. If you want to understand how individual position sizing interacts with portfolio-level drift, Dynamic Position Sizing covers the position-level decision that shapes what drift looks like in the first place.

A Worked Hypothetical: Applying a Policy to a Drifting Portfolio

All numbers below are invented. This is a design exercise, not advice about any real portfolio or allocation.

Suppose a hypothetical learner has designed a portfolio with four equal target weights of 25% each — positions A, B, C, and D. They choose a threshold policy: rebalance any position that drifts more than ±10 percentage points outside its 25% target. That means the tolerance band runs from 15% to 35% for each position.

Over some hypothetical period, Position A performs strongly. Its share of the total portfolio grows to 38%. Positions B, C, and D have each drifted proportionally lower. Under the threshold policy, Position A at 38% has crossed the 35% upper bound. The policy says: rebalance Position A back toward 25%.

Under a calendar policy, the same learner would have done nothing until the scheduled review date — regardless of whether Position A was at 38% or 28%. If the scheduled date had not yet arrived, the policy would dictate waiting, even though drift is substantial. Both responses are correct within their respective policies. They are not better or worse in absolute terms; they reflect different choices on the drift-cost-effort triangle.

Now apply the same drift scenario to the threshold policy with a tighter band: ±5 percentage points (band runs from 20% to 30%). In that case, Position A at 38% crossed the trigger much earlier in its drift — perhaps when it first reached 31%. The tighter band means more rebalancing events, more cost, and tighter drift control. The wider band means fewer events, lower cost, and more accepted drift.

The ±5% and ±10% figures used here are illustrative examples for the purpose of this exercise. They reflect ranges that have been discussed in the academic and practitioner rebalancing literature as reasonable design choices to examine. They are not recommended parameters for any real portfolio. Your appropriate band depends on your cost structure, how much drift you can tolerate, and how much monitoring effort you are prepared to commit to — none of which this article can determine for you.

The exercise also makes the observation-action link concrete. A portfolio review cadence — reading from the companion article at Portfolio Review Cadence — tells you when to look. The rebalancing policy tells you what to do when the look reveals a threshold has been crossed. Separating these two functions prevents both under-monitoring (missing drift because there is no scheduled look) and over-trading (rebalancing on every observation even when no threshold has been reached).

Risk Notes: What a Policy Cannot Do

A rebalancing policy does not prevent losses. It manages the allocation risk you are exposed to. If all your positions fall simultaneously — as happens in correlated sell-offs — rebalancing back toward target weights means buying positions that are declining. The policy is designed for that situation, but it does not protect against correlated drawdowns.

The relationship between correlation and concentration matters here. If the positions you are rebalancing among are highly correlated, your drift-control benefit is smaller than you might expect — when one falls, the others tend to fall with it, and the portfolio does not drift as much in the first place, but it also does not benefit from the natural rebalancing that comes from uncorrelated positions moving independently. Correlation and Concentration covers this in detail and is worth reading before designing a policy for any portfolio where the holdings share significant common risk.

A pre-committed policy also does not eliminate all behavioral pressure. You will experience the impulse to override it — especially when you are rebalancing by selling a position that "still looks strong" or by buying one that has fallen. The policy's value is precisely that it was written before you were in that position. Holding it requires recognizing that the act of renegotiating feels like good judgment but is structurally identical to the emotional drift the policy was designed to prevent.

Simulator Exercise: Write the Policy First, Then Watch It Work

Abu Terminal's simulator gives you a low-stakes environment to test a rebalancing policy before you have any real stake in whether it fires or not. The objective is behavioral, not financial: you are practicing the act of following a written rule under conditions that will generate pressure to override it.

Before starting a session, write down a complete rebalancing policy using this structure:

Policy type: Calendar or threshold (or hybrid).
Trigger condition: If calendar — specify the interval. If threshold — specify the band width, stated as a percentage-point deviation from target. Be exact; vague triggers will be renegotiated in the moment.
Action: When the trigger fires, what do you do? Write it as a single unambiguous sentence: "Return each out-of-band position to its target weight."
Hypothetical starting weights: Assign a target weight to each position in the simulator scenario (e.g., four equal 25% positions).

Run an Arena or Speed-Run session and observe how the positions drift across events. Apply your written policy mechanically — do not deviate from it based on how the moves feel. At the end of the session, record two things: how many rebalancing events your policy triggered, and the maximum drift from target that accumulated at any point.

Now run the same scenario under a different policy — tighter or looser band, or switch from threshold to calendar. Compare the number of rebalancing events and the maximum drift. The difference between those two numbers, across both runs, is the trade-off made visible.

The discipline being practiced is not which policy is better. It is the act of following the written rule without overriding it when the market appears to reward non-compliance. That is the behavioral skill this lab is building: mechanical execution of a pre-committed process, which is the same discipline required in any rules-based approach to managing risk.

Rebalancing Policy Lab