Rebalancing Methods

Strip away the jargon and rebalancing comes down to two decisions: when to do it and how to fund the trades.

When — The Trigger

Calendar

Rebalance on a fixed schedule, annually being the usual choice. Disciplined and trivial to automate, but blind to markets: it may trade when nothing has drifted and sit still through a violent move.

Threshold (banded)

Rebalance only when an asset’s weight strays past a set band — say $\pm 5$ percentage points, or a $\pm 25\%$ relative band. It reacts to what the market actually does and skips pointless trades, at the cost of needing the portfolio monitored.

Hybrid

Review on a calendar, trade only on a breach. This is the method the evidence supports: the discipline of a schedule, the cost control of a band, and neither overtrading nor unnoticed drift. Make it your default.

How — Funding the Trades For a taxable account this matters more than the trigger. Selling an appreciated holding to rebalance realizes a capital gain — a certain cost paid to correct a drift whose cost is only potential. Avoid it where you can:

• Rebalance with cash flows first. Steer dividends, interest, and new contributions into the underweight assets. In the accumulation years this alone can hold an allocation in line, with zero realized gains and zero spreads.
• Trade inside tax-advantaged accounts. Rebalancing within an IRA or 401(k) triggers no tax at all. Do the bulk of the selling there and leave the taxable account undisturbed.
• Harvest losses in tandem. When the taxable account must be traded, pair the rebalance with tax-loss harvesting (section “Tax-loss harvesting”) so realized losses offset the gains the rebalance creates.
• Upgrade while you are in there. When buying back an underweight class, ask whether a cheaper or better-built fund deserves the slot instead of the one you just sold.

For the few who automate execution through a broker API, the engineering checklist is short: budget for commissions and bid-ask spreads so they do not eat the benefit, use limit orders to control slippage, respect the wash-sale rule ( IRC §1091, “Loss from wash sales of stock or securities”) when harvesting, and keep a tax-aware ordering of accounts. Low-cost brokers with API access make this practical; the hard part is never the code, it is resisting the urge to trade more than the rules call for.

Rebalancing to a Moving Target Everything above assumes fixed target weights. Some frameworks let the target itself move. A risk parity rule sizes positions so each contributes equal risk, which means continuously trimming whatever has grown most volatile — useful when a portfolio holds assets of wildly different volatility. A factor tilt steers weight toward value, size, or quality exposures. These change what you rebalance to instead of how; they are asset-allocation decisions (section “Asset Allocation Models”) in a rebalancing costume, and they inherit every bit of the estimation-error fragility of the optimizers behind them.