A diversified portfolio drifts over time. Positions that perform well grow to larger shares of the portfolio; positions that perform less well shrink. Without rebalancing, a portfolio designed for a specific risk profile evolves into something with different risk characteristics — often more concentrated in what has recently performed well, which is precisely the wrong direction for someone worried about mean reversion.
Rebalancing brings the portfolio back to target weights, but the frequency and mechanism of rebalancing is a decision with meaningful long-run consequences. The three main approaches — annual, quarterly, and threshold-based — each have empirical trade-offs worth understanding.
The mechanical purpose
Rebalancing is fundamentally a mechanism for enforcing "buy low, sell high" without requiring judgment. When one asset class has outperformed another over a period, rebalancing sells some of the outperformer (which is now larger than target) and buys more of the underperformer (which is now smaller than target). The mechanic is countercyclical by construction: it fights against the natural drift toward whatever has recently done best.
The absence of rebalancing produces the opposite: a portfolio increasingly concentrated in whatever asset class has outperformed. Left unattended for a decade in most historical periods, a 60/40 stock/bond portfolio would drift meaningfully toward higher equity weights, dramatically shifting its risk profile in a direction the original allocation did not intend.
Annual rebalancing
The most common approach among retail investors and some institutional pension funds is annual rebalancing — checking positions once per year, typically in December or January, and returning them to target weights.
Empirical studies of the historical rebalancing frequency question generally find annual rebalancing produces slightly higher returns than quarterly rebalancing over long horizons, at the cost of somewhat higher intra-year volatility. The reason is that momentum tends to persist over multi-month horizons; a position that has outperformed for a few months often continues to outperform for a few more months. Rebalancing quarterly captures some of the mean reversion but also cuts off some of the ongoing momentum.
For taxable accounts, annual rebalancing has an additional advantage: rebalancing trades produce realised capital gains, and reducing trade frequency reduces realised-gain events. For tax-advantaged accounts, this consideration is irrelevant.
Quarterly rebalancing
Some pension funds and institutional accounts rebalance quarterly. The rationale is that shorter intervals produce more consistent risk exposure — the portfolio does not drift as far from target between rebalancings, so the risk profile stays more stable.
The trade-off is somewhat higher transaction costs, higher tax drag (in taxable accounts), and a slight give-up of momentum-driven returns in trending markets. Empirical studies find the return difference between annual and quarterly rebalancing is small (typically less than 30 basis points per year in either direction, depending on the specific decade studied). The choice is more about risk-profile stability than about return maximisation.
Threshold-based rebalancing
An alternative to calendar-based rebalancing is threshold-based rebalancing: check positions continuously (or on some frequent cadence), and rebalance only when a position has drifted beyond a specified threshold from target — typically 5% for major asset classes and 2–3% for smaller allocations.
The advantages of threshold rebalancing are that it captures more of the mean reversion benefit (rebalancing when the drift is largest) and that it minimises trades during periods of low volatility (when the drift is small). The disadvantage is that it produces variable turnover — some years will have several rebalance events, others will have none — which can complicate tax planning in taxable accounts.
Empirical studies generally find that threshold-based rebalancing produces slightly better risk-adjusted returns than pure calendar rebalancing over long horizons, but the difference is small enough that either approach is defensible.
The tolerance-band variant
A middle-ground approach combines calendar and threshold rules. Check the portfolio at scheduled intervals (say, quarterly), but only rebalance a position if it has drifted beyond a specified threshold (say, 5%) from target. Positions within the tolerance band are left alone; positions outside it are rebalanced.
This approach captures much of the risk-control benefit of frequent monitoring while reducing turnover and transaction costs. For most retail portfolios, the tolerance-band approach is a reasonable default: check quarterly, rebalance only when drift exceeds 5% for major asset classes.
The tax-loss harvesting connection
In taxable accounts, rebalancing decisions interact with tax-loss harvesting opportunities. If a rebalance would require selling a position at a gain, but a similar-exposure position is available at a loss, the tax-efficient version substitutes the loss-selling into the rebalance. Over time, this "tax-loss harvesting overlay" on the rebalancing framework can produce meaningful after-tax return improvements.
Getting this right is somewhat technical — wash sale rules, replacement securities that maintain similar exposure without triggering wash sale treatment, and the timing of realisations across tax years all matter — but the underlying principle is straightforward: rebalancing and tax-loss harvesting are complementary practices in taxable accounts.
The mechanics-versus-behaviour distinction
Beyond the specific mechanics, rebalancing has a subtler benefit for behavioural discipline. It forces the investor to sell some of what has recently performed well and buy some of what has not. This is the emotionally difficult direction — cutting winners is uncomfortable, adding to losers can feel like doubling down on a mistake — but it is the direction most portfolio strategies benefit from mechanically.
Investors who cannot bring themselves to rebalance during a strong period (because "why would I sell my winners?") end up with portfolios that drift toward concentration in whatever has done best. Investors who cannot rebalance during a weak period (because "I don't want to buy more of what's already down") end up under-allocated to whatever asset class has been beaten down. Both patterns produce worse long-run outcomes than mechanical rebalancing would have produced.
The rule to internalise
Rebalancing is not glamorous, but it is one of the most consistent contributors to long-run portfolio return quality across every academic study of the question. The specific frequency matters less than the discipline of doing it consistently — an annual, quarterly, or threshold-based approach will all outperform "no rebalancing at all" by meaningful margins over long horizons. The most important choice is not "which specific approach" but "which approach can you actually maintain for decades" — and the answer to that is usually the simplest, most mechanical version that requires the least ongoing discretion.
Educational content only. Not investment advice.