What does rebalancing actually do for you?
Rebalancing is selling some of what has risen and buying some of what has fallen, to return a portfolio to its target weights. It is widely sold as a way to increase returns. That claim is sometimes true, often false, and understanding when is what separates a real grasp of the mechanic from a slogan.
What it does reliably — risk control
Take the drifted portfolio from the last lesson: €112,800 sitting at 63.8/36.2 against a 60/40 target.
- Target equity value: 112,800 × 0.60 = €67,680
- Actual equity value: €72,000
- Trade: sell €4,320 of equities, buy €4,320 of bonds.
That single trade returns the portfolio to the exposure that was chosen. This is the effect rebalancing delivers every single time, in every market environment: it removes drift. Nothing about returns is required for this benefit to hold.
Be precise about what that buys, though. Rebalancing restores the weights, not the risk. Risk also depends on the volatilities and correlations behind those weights, and those move on their own — a 60/40 restored to 60/40 in a period of doubled equity volatility is carrying more risk than the same 60/40 did when the target was set. What rebalancing controls is the one part of that drift the owner actually causes.
What it does unreliably — returns
The "rebalancing bonus" is real but conditional. It arises when assets have similar long-run returns and high volatility around each other, so that periodic trades systematically sell high and buy low. Watch both cases with the same simple setup.
Case 1 — mean reversion (rebalancing wins). Asset A returns +30% then −20%; asset B returns 0% both years. Start €100 split 50/50.
- Unrebalanced: A ends at 50 × 1.30 × 0.80 = €52; B stays at €50. Total €102.
- Rebalanced after year 1: after year 1, A = €65, B = €50, total €115 → reset to €57.50 each. Year 2: A = 57.50 × 0.80 = €46, B = €57.50. Total €103.50.
Rebalancing added €1.50, by trimming A after it rose and before it fell.
Case 2 — a persistent trend (rebalancing loses). Asset A returns +10% every year for five years; asset B returns 0%. Start €100 split 50/50.
- Unrebalanced: A = 50 × 1.10⁵ ≈ €80.50; B = €50. Total ≈ €130.50.
- Rebalanced annually: the portfolio earns 5% every year → 100 × 1.05⁵ ≈ €127.60.
Rebalancing cost about €2.90, by repeatedly trimming the asset that kept winning.
The honest summary
Rebalancing is a risk-control mechanism with an ambiguous return effect. In markets that oscillate, it helps returns. In markets that trend, it hurts them. Nobody knows in advance which one the next decade will be, which is exactly why the risk-control justification is the durable one and the return justification is not.
You will see the return claim made confidently in marketing material. The two cases above are all you need to see through it.
The behavioural dimension
There is a further effect that no spreadsheet captures. Rebalancing requires buying the asset that has just fallen — in March 2009 or March 2020, that meant buying equities during a crash. A pre-committed rule makes that action mechanical rather than a fresh act of courage at the worst possible moment. Whether an individual can actually execute it is a question about them, not about the arithmetic, and it belongs to the capacity-versus-tolerance discussion from Unit 2.
Try it now
- Reproduce Case 1 and Case 2 above on paper. Confirm both numbers yourself — the point only lands once you have done the multiplication.
- Now find the day the rule would have fired. Navigate both charts to September 2008 – June 2009 and track the equity weight of a 60/40 across those months, Measuring each sleeve as you go. Find the first point at which equity drops below 55%. That is the day a band rule would have required buying equities. Note what the headlines were saying that week.
- State in one sentence what rebalancing guarantees and in one sentence what it does not. Keep both free of recommendations.