What rebalancing actually does to a portfolio
Rebalancing is the act of restoring your policy weights after markets have changed them for you. If you set 45% US equity and equities outrun bonds for two years, you now own more equity risk than you chose. Rebalancing sells a slice of what rose and buys what lagged, which returns the portfolio to the risk level your plan was built on.
The primary purpose is risk control, not return enhancement. A portfolio left alone for a long stretch of rising equity markets ends up dominated by equities, and the drawdown it experiences in the next bear market is the one you never agreed to. Whether rebalancing adds return depends entirely on whether asset classes mean-revert over your holding period, which is not something you can count on; whether it controls risk is arithmetic.
The calculator does three separate jobs. It measures drift, the gap in percentage points between each sleeve's current weight and its target. It computes the trades that close that gap exactly. And it solves the buy-only problem: the amount of new cash that would restore every target without a single sale, which matters enormously in a taxable account where a sale is a taxable event.
The three formulas, and why the trades always sum to your new cash
Current weight is each sleeve divided by the total: Vi ÷ ΣV. Drift is that weight minus the target weight, quoted in percentage points. A sleeve at 52.31% against a 45% target has drifted +7.31 pp. Percentage points are the right unit here; saying it has drifted “16%” (its relative overshoot) is a different measurement, and confusing the two is the most common error in rebalancing discussions.
The trade is the target value minus the current value: Ti = wi(ΣV + C) − Vi. Negative means sell, positive means buy. Because every target value is a share of the same total, the trades sum to Σwi(ΣV + C) − ΣVi = (ΣV + C) − ΣV = C. With no new cash they sum to zero, which is why the buys exactly fund themselves from the sells. That identity is the arithmetic check to run on any rebalancing plan.
The buy-only cash requirement is less obvious. If you may not sell, then no sleeve's value can fall, so the final portfolio total L must be large enough that every sleeve's current value already fits inside its target share: Vi ≤ wiL for all i. The smallest L that satisfies all four constraints is L = max(Vi ÷ wi), set by whichever sleeve is most overweight relative to its target. The cash you need is that L minus what you already have.
When your cash is less than that requirement, the calculator water-fills: it raises every underweight sleeve toward a common fill level until the cash runs out, so the most underweight sleeve is served first. That is the allocation that minimises the worst remaining shortfall, and it is what an advisor does by hand when a contribution lands mid-quarter.
Worked example: a $130,000 portfolio 7.3 points overweight in US equity
Take the calculator's defaults: US equity $68,000, international equity $22,000, bonds $35,000, cash $5,000, against targets of 45/20/30/5.
- Total the portfolio. 68,000 + 22,000 + 35,000 + 5,000 = $130,000.
- Find the current weights. 68,000 ÷ 130,000 = 52.3077%; 22,000 ÷ 130,000 = 16.9231%; 35,000 ÷ 130,000 = 26.9231%; 5,000 ÷ 130,000 = 3.8462%.
- Measure the drift. 52.3077 − 45 = +7.3077 pp; 16.9231 − 20 = −3.0769; 26.9231 − 30 = −3.0769; 3.8462 − 5 = −1.1538. The four drifts sum to zero, as they must.
- Compute the target values. 0.45 × 130,000 = $58,500; 0.20 × 130,000 = $26,000; 0.30 × 130,000 = $39,000; 0.05 × 130,000 = $6,500.
- Take the differences. Sell $9,500 of US equity; buy $4,000 international, $4,000 bonds and $1,500 cash. The three buys total $9,500, matching the sale exactly.
- Total the turnover. 9,500 + 4,000 + 4,000 + 1,500 = $19,000 changes hands, 14.6% of the portfolio.
- Price the buy-only alternative. The binding constraint is US equity: 68,000 ÷ 0.45 = $151,111.11. Subtract the current $130,000 and you need $21,111.11 of new money to reach every target without selling anything.
That last figure is the one worth staring at. Fixing a 7.31 point drift with trades costs $19,000 of turnover and, in a taxable account, a realised gain on the $9,500 sale. Fixing it with contributions costs no tax at all but requires $21,111.11 — 16.2% of the portfolio — which is why buy-only rebalancing works well early in an accumulation phase and stops being practical once the portfolio is large relative to your annual savings.
How wide should your tolerance band be?
Two policies are in common use and both are defensible. Calendar rebalancing means trading on a fixed schedule — annually is the usual choice — regardless of drift. It is simple, it is easy to automate, and it guarantees you will sometimes trade when nothing needed fixing.
Band rebalancing means trading only when a sleeve breaches a tolerance. The widely used version is the 5/25 rule: act when a sleeve moves 5 absolute percentage points from target, or 25% of its own target weight, whichever is smaller. The two triggers cross over at a 20% target weight, because 25% of 20 is exactly 5. Above 20% the absolute rule binds; below it the relative rule binds, which is what stops a 5% satellite position from being allowed to double before anyone notices.
A narrow band controls risk more tightly and costs more in trading and tax. A wide band does the reverse. In a tax-deferred account where trades are free, a tight band is close to costless and there is little reason not to use one. In a taxable account, every sale of an appreciated position triggers a bill, so the sensible order of operations is: direct new contributions to the underweight sleeves first, then use dividends and interest, then let the band do its work, and only sell when nothing else closes the gap. Feed any sale you do make through the capital gains tax calculator before you place it.
Read the drift signs together rather than one at a time. Because the four weights must sum to 100%, an overweight in one sleeve is necessarily an underweight somewhere else. In the worked example a single +7.31 point overweight is spread across three underweights of 3.08, 3.08 and 1.15 points, none of which breaches a 5-point band on its own. A band policy applied sleeve by sleeve therefore fires on the overweight and stays silent on the underweights, which is the correct behaviour: fixing the one triggers the others automatically.
Where the 5/25 rule triggers, by target weight
| Target weight | Absolute trigger | Relative trigger (25%) | Which one binds | Acts at a weight of |
|---|---|---|---|---|
| 5% | 5.00 pp | 1.25 pp | Relative | 3.75% or 6.25% |
| 10% | 5.00 pp | 2.50 pp | Relative | 7.50% or 12.50% |
| 15% | 5.00 pp | 3.75 pp | Relative | 11.25% or 18.75% |
| 20% | 5.00 pp | 5.00 pp | Identical | 15.00% or 25.00% |
| 30% | 5.00 pp | 7.50 pp | Absolute | 25.00% or 35.00% |
| 45% | 5.00 pp | 11.25 pp | Absolute | 40.00% or 50.00% |
| 60% | 5.00 pp | 15.00 pp | Absolute | 55.00% or 65.00% |
The crossover sits at a 20% target because 0.25 × 20 = 5. This calculator applies a single absolute band; set it to 0.25 × your smallest target weight if you want the relative rule to govern a small satellite sleeve.
Mistakes that make a rebalancing plan wrong
- Rebalancing account by account instead of across the whole portfolio. Your asset allocation is a property of everything you own. Rebalance at the household level and place each asset class in whichever account taxes it most lightly.
- Confusing percentage points with percent. A sleeve at 52.31% against a 45% target has drifted 7.31 percentage points, which is a 16.2% relative overshoot. Bands are quoted in different units by different sources, so check which one you are reading.
- Selling appreciated positions in a taxable account first. Exhaust new cash, dividends and interest before you sell. The buy-only figure above tells you how far that route can take you.
- Rebalancing too often. Every trade has a cost, and drift within your band is not a problem — it is the tolerance working as designed.
- Ignoring wash-sale rules when harvesting losses at the same time. Buying a substantially identical security within the 30-day window on either side of a loss sale disallows the loss.
- Treating a target weight as a forecast. The weights encode the risk you chose, not a view about which asset class will win next year. Changing them because one sleeve is doing well is not rebalancing; it is a strategy change wearing a rebalancing costume.
Rebalancing inside a larger plan
Rebalancing sits between two other decisions. Above it is the allocation itself — the target weights, which should come from your horizon and your capacity to tolerate a drawdown, not from recent performance. Below it is the mechanics of getting money in, which for most people is a monthly contribution: if that contribution is directed at whichever sleeve is furthest below target, the portfolio largely rebalances itself for years. The SIP calculator and the dollar-cost averaging calculator model the contribution side.
As a portfolio grows, the balance of power shifts. Early on, contributions dwarf drift and buy-only rebalancing handles everything. Later, a single year's market move can be larger than a year's savings, and trading becomes unavoidable. The buy-only cash figure above is the exact point at which that transition happens for your portfolio: when it exceeds what you can realistically add, you are in the trading regime.
In retirement the same machinery runs in reverse. Withdrawals are the natural rebalancing tool: take the money from whichever sleeve is most overweight and the allocation is maintained at zero extra cost. That is a free improvement over selling proportionally, and it pairs directly with the spending framework in the safe withdrawal rate calculator and the target set by the FIRE number calculator. If you are adding to a single position rather than an asset class, the average down calculator handles the cost-basis side of the same problem.