Pension Rebalancing at Quarter-End: How the 60/40 Trade Works and Who Front-Runs It

In 2025 equity mutual funds had $1.2 trillion of net outflows, and the ICI Fact Book named one cause in plain terms: stocks beat bonds by a wide margin, so investors and target-date funds following allocation rules had to sell equity funds and buy bond funds to stay on target. That is rebalancing, and it is one of the few large flows in markets whose direction you can work out from public prices before it happens. A 2025 NBER paper estimates it costs investors about $16 billion a year, partly because other traders see it coming. If the idea of mechanical flows is new, start with the hub on institutional money flow.

What is pension rebalancing at quarter-end?

It is the trading that funds with fixed asset-allocation targets do around the end of each quarter to restore their stock and bond weights. Pensions, sovereign funds, target-date funds and balanced mutual funds all do some version of it.

A fund that wants 60% in stocks and 40% in bonds will not stay there. If stocks rally, the equity weight rises; if they fall, it shrinks. Rebalancing sells the winner and buys the loser to get back to target. Harvey, Mazzoleni and Melone (NBER w33554) note that large pensions and sovereign funds commonly target a 60/40 mix, and that because these funds have relatively tight mandates and stable asset shares, they must buy losers and sell winners as prices move. The trade is contrarian by design, and that is what makes it predictable.

Why do 60/40 funds rebalance, and how do the rules differ?

Funds rebalance to keep the risk they chose. Left alone, a 60/40 portfolio after a long bull market can become 75/25 and carry far more equity risk than its board approved. The two main rule types are calendar and threshold.

• Calendar rebalancing trades back to target on fixed dates, usually month-end or quarter-end. Simple, auditable and predictable.
• Threshold (band) rebalancing trades only when a weight drifts outside a permitted range around target. It trades less often but can sit some distance from target.
• Hybrids check the bands on a schedule, so they trade on calendar dates but only when drift is big enough.

Real policies use ranges. The CalPERS Total Fund Investment Policy, effective June 10, 2024, sets these targets and ranges:

The policy requires “adherence to the asset class policy ranges” with rebalancing done efficiently and prudently, and its policy benchmark is rebalanced quarterly. Wide bands like these give staff discretion over timing, which is one reason pension trades are spread out more than the headlines about a single “quarter-end dump” imply. For how targets and ranges are set in the first place, see strategic vs tactical asset allocation.

When does rebalancing happen: month-end or quarter-end?

Both, but the pressure is strongest in the last days of the month and builds into quarter-end. Month-end trades are tied to cash needs; quarter-end trades add the calendar rebalancers and reporting dates.

Harvey, Mazzoleni and Melone point out that mature pension funds often sell assets at month-end to raise cash for benefit payments, and that many institutions rebalance monthly or quarterly. In their data, the calendar-based signal predicts returns strongly at month-end and not at other times, and the predictive power of both their calendar and threshold signals increases as quarter-end approaches. December stacks everything at once: quarter-end, year-end reporting, tax trades and, historically, window dressing.

How do you estimate the quarter-end rebalancing trade?

Use trade ≈ w × (1 − w) × (stock return − bond return) ÷ (1 + portfolio return), where w is the equity target. The result is the share of the fund that must move from the winner to the loser.

Worked example, illustrative. A $100 billion fund starts the quarter at 60/40. Stocks return +10% and bonds −2%.

1. Stocks grow from $60.0B to $66.0B; bonds fall from $40.0B to $39.2B.
2. The fund is now $105.2B, and the equity weight is 66.0 ÷ 105.2 = 62.7%.
3. The target equity holding is 60% × $105.2B = $63.12B.
4. The fund must sell $66.0B − $63.12B = $2.88B of stocks and buy the same amount of bonds.
5. That is 2.7% of the fund. The formula gives the same answer: 0.6 × 0.4 × 12% ÷ 1.052 = 2.74%.

Three adjustments make the estimate more realistic. Bands mean some funds won’t trade until drift exceeds their range. Contributions and benefit payments let funds rebalance with cash instead of selling. And many funds have already traded partway through the quarter. So treat the formula as the direction and rough upper bound, not a precise order size.

How big has the 60/40 rebalancing trade been?

For a quarterly-rebalanced 60/40 portfolio, the implied trade has usually been a few percent of the fund, mostly selling stocks, with large buy-stock trades after crash quarters.

We ran a simple 60/40 portfolio through ten years of FRED data, using the S&P 500 price index for stocks and a duration-based 10-year Treasury proxy for bonds. In 33 of 40 quarters from Q4 2016 to Q3 2026, the quarter-end trade was to sell stocks, because stocks usually outran bonds. The exceptions are the stress quarters. After Q1 2020 the equity weight had fallen to about 52%, so the fund needed to buy stocks worth about 7.9% of the portfolio. Q4 2018 called for about 4.5% and Q2 2022 for about 3.2%, the year stocks and bonds fell together, a regime covered in stock-bond correlation. The quarter after the 2020 crash flipped to selling about 4.2% as stocks recovered.

That pattern matters because it is countercyclical: rebalancers buy after selloffs and sell after rallies. It works against the forced sellers you meet elsewhere in this series, such as volatility control funds and trend followers tracked through CTA positioning, which sell into weakness. When both groups are large, quarter-end can be a tug of war.

Do traders front-run pension rebalancing?

The evidence says yes, and it costs the rebalancers. Because the trade is predictable from public returns, others can position ahead of it.

Harvey, Mazzoleni and Melone find that when stocks are overweight, funds sell stocks and buy bonds, and equity returns are 17 basis points lower over the next day, a result robust to momentum, reversals and macro controls. They estimate rebalancing costs institutions more than 8 basis points a year, about $16 billion annually or roughly $200 per US household, which they note can exceed the fees those funds pay to invest passively. A strategy that trades S&P 500 futures against 10-year Treasury futures ahead of the predicted rebalancing earned a Sharpe ratio above 1 from 1997 to 2023 in their tests. They also show the predictability became significant in the early 2000s.

Calendar vs threshold rebalancing: which is better?

For the fund, the less predictable rule usually costs less. Calendar rules are easiest to front-run; bands and randomized timing are harder to anticipate.

The NBER authors simulated rebalancing on randomized dates with the same average frequency as a monthly calendar and found an average cost of about 0.6 basis points, which they describe as economically negligible. The lesson for you is the same as for a pension: if many people must trade at the same moment, don’t be one of them. Spreading trades over several days, or using bands, keeps most of the risk control without paying the crowd’s price. The SEC’s asset allocation guide describes both calendar and threshold approaches for individual investors.

What is window dressing at quarter-end?

Window dressing is reshaping a portfolio just before a reporting date so it looks better, typically by dumping losers so they don’t appear in the quarter-end holdings list.

The classic evidence is a 1991 study by Lakonishok, Shleifer, Thaler and Vishny (NBER w3617) of 769 pension funds with about $129 billion in assets. Managers oversold poorly performing stocks, and those sales increased in the fourth quarter, when sponsors review portfolios most closely. It is older evidence, and reporting and benchmarking have changed since, so treat window dressing as a modest year-end effect on recent losers rather than a market-wide flow. The quarter-end holdings it targets are the ones that later appear in 13F filings.

What does quarter-end rebalancing mean if you hold concentrated stock?

If you are an executive or tech employee with a large single-stock position, rebalancing flows are a timing input, not a trading signal. The goal is to avoid selling or hedging into the most crowded windows.

After a strong quarter for stocks, broad selling pressure into quarter-end can weigh on large index constituents, including your company. If your diversification sales run through a 10b5-1 plan, schedule tranches across the month rather than on the last trading days. If you are adding a collar, a pullback driven by mechanical selling can be a better time to buy the put than after a rebound. Company-specific windows matter too: the buyback blackout period before earnings often overlaps quarter-end, removing the corporate bid exactly when rebalancers sell. When you reinvest the proceeds into ETFs, their in-kind structure, explained in our guide to ETF fund flows, keeps unwanted capital gains distributions down.

The same arithmetic applies to your own plan. Your concentrated position is a portfolio that has drifted far from any target. Rebalancing it on a schedule you set in advance, rather than reacting to quarter-end moves, is the institutional discipline worth copying.

What is a checklist for quarter-end rebalancing?

1. Track the quarter-to-date gap between stock and bond returns.
2. Estimate the drift with w(1−w)(gap) ÷ (1 + portfolio return); note the sign.
3. Flag the last five trading days of the quarter and of each month on your calendar.
4. Check what else is forced that week: buyback blackouts, option expiration, index rebalances.
5. Spread your own large sales or hedges across days, away from the final close.
6. Expect the effect to be small and short-lived; don’t trade it with leverage.
7. Rebalance your own concentrated position by a written plan, not by headlines.

Frequently asked questions

Key takeaways

1. Fixed-allocation funds must sell the asset that outperformed and buy the one that lagged; the direction is knowable from public prices.
2. Estimate the trade as w(1−w)(stock return − bond return) ÷ (1 + portfolio return).
3. In a FRED-based 60/40 simulation, 33 of 40 quarters from Q4 2016 to Q3 2026 called for selling stocks; Q1 2020 called for buying about 7.9% of the fund.
4. Harvey, Mazzoleni and Melone tie rebalancing to a 17 basis point next-day drag on stocks and about $16 billion a year in costs.
5. Month-end trades link to pension cash needs; predictability rises into quarter-end.
6. Window dressing is documented but older evidence; treat it as a small year-end effect.
7. If you sell or hedge a concentrated position, avoid placing size into the most crowded rebalancing closes.

Sources and method

1. Harvey, Mazzoleni and Melone, The Unintended Consequences of Rebalancing (NBER w33554, March 2025, revised January 2026) — 17 bp next-day effect, $16 billion annual cost, month-end and quarter-end timing, front-running strategy, randomization
2. CalPERS Total Fund Investment Policy (effective June 10, 2024) — asset class targets and ranges; quarterly rebalanced policy benchmark
3. Lakonishok, Shleifer, Thaler and Vishny, Window Dressing by Pension Fund Managers (NBER w3617, 1991) — selling of losers rises in the fourth quarter
4. ICI, 2026 Investment Company Fact Book, chapter 3 — rebalancing as a driver of 2025 equity mutual fund outflows
5. S&P 500 via FRED (SP500) — equity leg of the 60/40 simulation, October 2016 to September 2026
6. 10-year Treasury yield via FRED (DGS10) — bond proxy for the 60/40 simulation
7. SEC, Beginners’ Guide to Asset Allocation, Diversification, and Rebalancing — calendar and threshold rebalancing basics

Chart 2 is a Proflex simulation, illustrative: a 60/40 portfolio of the S&P 500 price index (FRED SP500, no dividends) and a 10-year Treasury proxy whose daily return is the prior yield divided by 252 minus 8.0 times the change in yield (FRED DGS10), rebalanced on the last trading day of each quarter from Q4 2016 to Q3 2026. Omitting dividends understates equity drift by roughly the dividend yield. Charts 1 and 3 are Proflex frameworks. The worked example uses round hypothetical returns.