Rebalancing
Periodic and drift rebalancing, and how each changes the results.
The weights you set apply on day one. From then on markets move and the allocation changes by itself, because whatever returns most weighs more and more on the total. After a few years a portfolio can carry a very different risk profile from the one you chose.
Rebalancing means selling what has grown beyond target and buying what has lagged, bringing weights back to their set values. The goal is to keep risk within the limits you decided at the start. The effect on return depends on the period observed and can be positive or negative: it is not a technique for earning more.
Three rebalancing rules applied to the same price series. The series is simulated and only illustrates the mechanism: it does not correspond to any real historical period. With no rebalancing the equity share moves from 60% to 74%, leaving the portfolio more exposed than intended. Annual rebalancing returns it to 60% every January, for ten interventions over ten years. Drift at 5 points keeps it inside the 55%-65% band throughout, with five interventions instead of ten.
The same portfolio shown as composition. This is the chart available under “Composition” after running the analysis. The “R” lines mark the months when a rebalance was triggered.
The three modes
Frequency is chosen in the settings:
| Option | Behaviour |
|---|---|
| Monthly, Quarterly, Semi-annual, Annual | Rebalances on a fixed schedule |
| Drift-based | Rebalances when an instrument moves further from target than the set threshold |
| Never | Buy & hold: weights evolve freely |
More frequent rebalancing keeps the allocation closer to target, but means more transactions. If you enabled capital gains tax and commissions, their cost is deducted from results, which is exactly why "rebalance as often as possible" isn't the right answer.
How drift works
Drift thinks in bands rather than dates. With a 5-point threshold on a 60% target, the band runs from 55% to 65%.
Weights are checked every month: while they stay inside the band nothing happens; as soon as even one instrument leaves it, the whole portfolio returns to target weights.
The number of interventions isn't decided in advance, it depends on how much markets move. In a calm period years can pass without a single trade; in a turbulent one several cluster together. A tight threshold behaves like very frequent rebalancing, a wide one behaves like Never.
The threshold is absolute
It is percentage points of the portfolio total: target 60%, current weight 54% → 6 points of drift.
Examples
- 2 instruments, target 60/40, threshold 5 → weights 68/32, 8 points of drift: outside the band, full rebalance to 60/40.
- 2 instruments, same target → weights 63/37, 3 points: nobody breaches the threshold, no trades.
- 3 instruments, target 60/20/20, threshold 5 → weights 54/23/23: A is 6 points outside its band, so a full rebalance to 60/20/20, even though B and C were still inside their bands.
- With a DCA plan, if drift triggers in a contribution month, the instalment and the rebalance happen together: portfolio plus instalment are brought back to target weights. In other months the instalment is simply invested according to target weights.
With more than two positions
The logic is unchanged, but read instrument by instrument. Each position has its own band around its own target: with a 5-point threshold, a 40% holding has a 35%–45% band, a 10% holding has 5%–15%.
Every month each instrument is checked against its own band, and one breach is enough to bring the entire portfolio back to target.
The caveat that matters. A threshold in absolute points doesn't weigh equally across rows: the same 5 points are tight on large positions and loose on small ones. A 40% position leaves its band after losing roughly 12% of its own relative weight; a 10% position has to reach one and a half times its target before it does.
So if the portfolio holds many small positions, use a lower threshold, otherwise individual rows can move a long way without triggering anything.
Which one to choose
There's no universal answer, but there is a practical criterion: the right mode is the one that keeps risk where you want it at the lowest cost.
- Annual is the sensible default for most portfolios: few interventions, contained drift, predictable costs.
- Drift acts only when it's genuinely needed, and does far less than a calendar in calm periods. It makes sense if commissions bite, or if you prefer a rule tied to risk rather than to the date.
- Never is a legitimate choice only if you accept that the portfolio becomes progressively riskier. Over ten years an unrebalanced 60/40 tends to end up around 74% equity.