The VIX spent much of the first half of 2026 grinding lower before geopolitics reasserted itself. On July 13, the index surged to an intraday high of 17.40 as U.S.-Iran hostilities flared over the Strait of Hormuz, sending the Nasdaq down 1.5% and Brent crude up nearly 10% in a single session — a sharp move, but still well below the peaks above 35 the VIX touched during the initial Iran-U.S. escalation in February and March. That pattern — long stretches of subdued realized volatility interrupted by sharp, headline-driven spikes — is precisely the environment in which the mechanics of rebalancing, not just the decision to do it, start to matter for portfolio outcomes.
Threshold vs. Calendar: A Settled Debate, Mostly
Calendar rebalancing — quarterly, annual, whatever cadence a policy statement specifies — has the virtue of simplicity and auditability. Threshold (or “tolerance band”) rebalancing instead lets allocations drift until a position breaches a defined band around target weight, then trades only that position back toward target. The research consensus, built over two decades, favors thresholds, though the margin is narrower than practitioners sometimes assume. Vanguard’s classic study by Jaconetti, Kinniry, and Zilbering, running a 60/40 portfolio back to 1926, found essentially no difference in risk or return between monthly and annual calendar rebalancing — the only reliable effect of higher frequency was more turnover. The more decisive evidence for thresholds comes from Gobind Daryanani’s 2007 Journal of Financial Planning study, which tested rolling five-year windows from 1992 to 2004 on a five-asset-class 60/40 portfolio and found the optimal relative band was roughly 20% of a position’s target weight — narrower bands (10-15%) underperformed by trading too early, wider bands (25%+) underperformed by trading too late, and checking for breaches less often than every ten trading days measurably eroded the benefit. Vanguard’s more recent target-date fund research extends this into a hybrid “200/175” design — monitor allocations continuously, rebalance only when a sleeve drifts 200 basis points from target, and reset it back to within 175 basis points rather than exactly to target. That structure is estimated to add roughly 15-22 basis points annually during accumulation and 22-25 basis points during retirement drawdowns versus calendar rebalancing, while generating only about a quarter of the transaction costs of monthly rebalancing.
What the Return Studies Actually Show
Advanced investors should be precise about what this literature is and isn’t claiming. Rebalancing is not primarily a return-enhancement strategy — it is a risk-control discipline that occasionally, incidentally, pays a modest premium. Morningstar’s long-running test of a 60/40 (S&P 500 / Bloomberg U.S. Aggregate) portfolio since 1994 is instructive: during the COVID drawdown, an un-rebalanced portfolio that had drifted to roughly 80% equity lost 27.8% between February 19 and March 23, 2020, versus 20.7% to 21.3% for portfolios held to any rebalancing discipline. Over the trailing 15 years through mid-2020, annual rebalancing delivered the lowest standard deviation (8.55%) and the lowest downside-capture ratio (54.12%) of any tested frequency, while a 5%-band threshold strategy edged out the others on upside capture. Morningstar’s own conclusion is blunt: the differences between reasonable rebalancing frequencies “aren’t really big enough to lose any sleep over.” The real cost of ignoring rebalancing shows up not in average years but in the tails — the un-rebalanced portfolio’s equity weight had drifted to nearly 80% heading into two of the three worst drawdowns tested precisely because winners had been left to run.
A Correlation Regime That Won't Sit Still
The case for tolerance-band discipline gets more urgent when the diversification benefit driving the whole exercise is itself unstable. Vanguard’s own analysis puts the average global stock/bond correlation at roughly -7% for 2000-2021, versus +33% in the 1990s, and notes that 2022 was the first calendar year since 1977 in which both equities and bonds posted simultaneous losses — before the negative relationship reasserted itself in 2023. Where that correlation goes next is genuinely contested among institutional strategists. PIMCO’s analysis of 2000-2024 data found that 60/40 drawdowns have historically been shallower when the stock/bond correlation is negative (median monthly loss of -1.5%) than when it’s positive (-2.1%), and argued as inflation and growth moderated the negative relationship had reasserted itself. Oxford Economics takes the opposite side, forecasting a return to positive stock/bond correlation through 2026 on the back of persistent supply shocks, elevated term premia, and more activist fiscal policy — conditions that, in its framework, undermine the assumption that government bonds reliably cushion equity drawdowns. For a rebalancing program, the practical implication of this disagreement is not to guess which forecast wins, but to recognize that a static band calibrated during the 2010s low-correlation, low-vol regime may now be miscalibrated in both directions at once: elevated volatility means individual sleeves move further and faster, breaching fixed-percentage bands more often and generating more turnover than the original calibration assumed, while a less reliably negative correlation means the diversification payoff from each rebalancing trade is smaller than it used to be. That argues for periodically re-stress-testing band widths against trailing realized volatility rather than treating a 5% or 20%-relative threshold as permanent.
Transaction Costs and Tax Mechanics
For taxable accounts, the theoretical rebalancing premium is easily overwhelmed by frictions if execution is careless. The Daryanani framework’s core practical advantage — trading only the sleeve that breaches its band rather than touching every position on a calendar date — is what drives Vanguard’s finding that threshold rebalancing costs roughly a quarter of what monthly calendar rebalancing costs. Beyond that structural saving, three practices do the most work for advanced investors managing taxable capital. First, use new contributions, dividends, and distributions to close drift before ever selling an appreciated position — cash flow rebalancing costs nothing in realized gains and is underused relative to its impact. Second, sequence rebalancing trades by account: execute the sale of an overweight, appreciated equity sleeve inside the IRA or 401(k) first, and reserve the taxable account for trimming positions with embedded losses or for tax-loss harvesting pairs, since 2022’s bond repricing left many taxable fixed-income holdings sitting on unrealized losses that can be harvested opportunistically alongside a threshold breach. Third, where the goal is simply to trim equity beta without triggering a gain, index futures, options overlays, or completion portfolios can adjust net exposure synthetically, deferring the taxable event while still restoring the target risk profile — a technique increasingly available to non-institutional investors through options-based ETF structures.
None of this argues for abandoning discipline in favor of forecasting. It argues for treating the tolerance band itself as a parameter worth revisiting — not because volatility changes the case for rebalancing, but because it changes the cost of getting the band width wrong.