Most investing books say "rebalance annually." For a 401k that is fine. For a taxable brokerage it is the single most expensive habit you can keep — and it leaves 0.5–1% of your portfolio on the table every year, year after year.
The Question Everyone Gets Wrong
Search "how often should you rebalance your portfolio" and the top results will tell you once a year, or once every six months, or "when your allocation drifts more than 5 points." All of those answers assume one thing: that rebalancing is free.
It isn't. Every rebalance in a taxable account is a sale. Every sale is a realization event. Every realization event is a tax bill. If you rebalance once a year without an offsetting strategy, you systematically convert unrealized paper gains into realized taxable gains — and you forfeit the tax-loss harvesting you'd otherwise collect from the same drift.
The right schedule depends on the account. In a 401k or IRA, rebalance mechanically: cheap, fast, no tax consequence. In a taxable brokerage, the schedule is determined by how much tax you can avoid, not by how clean your allocation looks.
The rule: In a tax-advantaged account, rebalance on cadence (calendar or threshold). In a taxable account, rebalance on cadence AND pair every rebalance with tax-loss harvesting — or you are paying the IRS to keep your allocation neat.
Why Frequency Matters More in Taxable Accounts
Three forces make rebalancing frequency a tax story, not an allocation story, the moment any of your money sits in a taxable brokerage:
1. Drift triggers realization, harvesting captures offset
A 60/40 portfolio drifting to 72/28 over 12 months is not just an allocation problem. The position that grew (say U.S. equity) is carrying an unrealized gain. The position that shrank (say international equity) is carrying an unrealized loss. When you rebalance, you sell some of the winner and buy more of the loser — realizing the gain and absorbing the loser at a higher cost basis. Net result: tax bill, no offset.
If you harvest as you rebalance, the picture inverts. Sell the underwater position into a similar (not "substantially identical") replacement, lock in the loss, then buy your target allocation with the proceeds. Same drift, same target — but now you have losses that offset future gains, plus up to $3,000/year against ordinary income.
2. Higher frequency sweeps up more losses
Markets don't lose money once a year. Individual positions drift in and out of loss territory all year long. An annual rebalance sees only the year-end snapshot. A quarterly sweep sees four. Continuous or weekly scanning captures much more.
The empirical alpha from pairing frequent rebalancing with TLH is meaningful: published direct-indexing manager results (where monitoring is daily) commonly show 0.8–1.2% of pre-tax alpha over a buy-and-hold benchmark after tax, of which roughly half comes from the harvest sweep itself and half from keeping the allocation tighter.
3. Long compounding multiplies small differences
A 0.8% annual edge looks small until you compound it. Over 20 years on a $250,000 portfolio that grows to roughly $1.1M assuming 8% gross, that 0.8% tax-difference compounded is the difference between $1.05M and $1.13M after-tax at year 20 — a gap of about $80,000 from a single schedule choice.
The Rebalancing Schedules (And When to Use Each)
There are three standard schedules. They are not mutually exclusive — most sophisticated investors run one primary and one trigger-based overlay.
| Schedule | How It Works | Best For | Tax Cost in a Taxable Account |
|---|---|---|---|
| Calendar (annual) | Rebalance on a fixed date once a year (Dec, your birthday, etc.). | 401k, IRA, simple portfolios, hands-off investors. | High — one year of drift crystallized into gains, no harvesting. |
| Calendar + TLH (quarterly) | Rebalance every 3 months; sweep losses with each pass. | Taxable brokerage with $100K+; most sensible default. | Low to moderate — losses partially offset gains each quarter. |
| Threshold (5/25) | Rebalance when any asset class drifts ±5 absolute percentage points or ±25% relative. | Volatile allocations (small-cap, EM, alts); supplements calendar. | Variable — fires when drift matters, but can over-trade in choppy markets. |
| Continuous (real-time) | Daily or weekly monitoring; harvest on loss, rebalance on drift. | Direct-indexing mandates, sophisticated retail with tooling. | Lowest — captures nearly all loss opportunities, washes less drift into gains. |
A $250,000 Worked Example: Annual vs. Quarterly-Plus-TLH
Suppose you hold a $250,000 portfolio targeting 60% U.S. equity / 40% international equity in a taxable brokerage. Markets are good this year — U.S. outperforms. After 12 months, your portfolio has drifted to roughly 72% U.S. / 28% international, a $30K overweight in U.S. equity. International has lagged and several positions sit on paper losses.
Drift: 60/40 → 72/28 over 12 months (U.S. equity outperformed).
Overweight sale: Sell $30,000 of U.S. equity to bring allocation back to 60/40. Average unrealized gain on the sold shares: ~$14,000 (long-term).
Tax bill: $14,000 × 15% LTCG = $2,100 federal long-term capital gains tax. Plus state tax in most states — call it $2,400 total.
Harvested losses: $0. The underwater international position is held, not sold — so the loss is never realized.
Net after-tax drag from rebalancing alone: ~$2,400.
Drift: Same 60/40 → 72/28 drift over 12 months.
Each quarter: Sweep the portfolio for positions sitting on losses. Sell each loss position into a similar-but-not-substantially-identical replacement (e.g. swap IXUS → VXUS, VTI → ITOT) to stay invested while locking the loss. Total losses harvested across the year: ~$11,000.
Quarterly rebalance trim: When the 60/40 allocation drifts past threshold, trim the overweight — the gain realized is much smaller because the trim is partial and frequent. Total gains realized: ~$5,000.
Tax math: $5,000 gains realized, $11,000 losses realized — net $6,000 of long-term losses offsetting other gains, plus $3,000 against ordinary income.
Net federal tax bill: $0 (losses exceed gains) and a $2,820 reduction in ordinary income tax at a 24% bracket ($3,000 × 0.94 effective rate). Plus state savings on the $3,000 ordinary offset.
Tax savings vs. Plan A: ~$3,800 in the same year on the same drift.
Two plans on the same $250,000 portfolio, the same drift, the same year. Plan A leaves $3,800 on the table. Compounded over 20 years that number is large enough to fund a meaningful chunk of retirement — and the gap accrues every single year you keep rebalancing without harvesting.
Practical Rules to Implement This
Three rules cover most situations for a self-directed taxable portfolio:
- Pick a primary cadence: quarterly is the sensible default for most taxable portfolios above $50K. Below $50K, losses harvested are too small to justify transactional costs — annual is fine.
- Pair every rebalance with a loss sweep: before selling any winner, scan for losers and lock them first. The math above only works if the sweep happens in the same transaction window.
- Add a drift overlay: if any single position drifts more than ±5 percentage points absolute, harvest it immediately regardless of where you are in your quarterly cycle. Drift binds before cadence.
The shortcut: scan your portfolio free with WealthPilotOS. We flag lost positions, drift, and wash sale risk every day — the underlying inputs that let your quarterly cycle actually generate the missing tax savings.
What About Threshold-Only Rebalancing?
Threshold-only (rebalance only when an asset class drifts past ±5/+25%) reduces transaction count, but it has a real downside in a taxable account: when threshold triggers, the drift is already large, so the trim is also large. Large trims realize large gains. You can layer TLH onto a pure-threshold strategy, but a quarterly cadence sweeps more loss opportunities than a trigger-driven approach — direct-indexing empirical data backs this up.
The mixed schedule — quarterly cadence with a 5/25 drift overlay — is what most institutional direct-indexing mandates run, and it is the lowest-tax-cost schedule accessible to retail accounts today.
The Schedule That Works
In a 401k or IRA: pick whatever cadence keeps you disciplined. Calendar or threshold — it does not matter because the IRS does not see any of it.
In a taxable brokerage: quarterly calendar plus daily loss harvesting plus a drift overlay. The combined schedule is what closes the 0.5–1% gap that an annual-only rebalance leaves behind every year.
If you have a taxable account today, run it through WealthPilotOS — see exactly how much you would have harvested with this schedule over the past year, and what the gap is between where you are and where the schedule would put you.