NextFin News - Every quarter, like clockwork, more than $40 billion slides out of BlackRock's iShares Core S&P 500 ETF and into a near-identical fund run by Vanguard — only to reverse course a few days later. The money is not chasing performance. It is dodging a tax: the 30 percent US withholding levy on dividends paid to foreign investors, made almost mechanical by the two asset managers' misaligned dividend calendars.
The pattern is large enough to stand out even in a $16 trillion US exchange-traded fund market, and regular enough that its next appearance can be predicted with confidence: the middle of September. It is a quiet arbitrage built on a simple fact of market plumbing. BlackRock's IVV and Vanguard's VOO offer the same S&P 500 exposure, but since 2024 BlackRock's ex-dividend dates have consistently fallen earlier than Vanguard's. Large overseas holders sell IVV before its shares go ex-dividend, park the proceeds in VOO, wait out BlackRock's record date, then switch back into IVV before Vanguard's own ex-dividend date arrives. At every moment they are fully invested in the index. At no moment are they on either fund's books when a dividend is declared payable.
Calculations based on the dollar value shifting between the two funds each quarter suggest the IVV-VOO flip saved foreign investors roughly $147 million in US taxes last year. The practice appears legal, and Treasury officials have signaled it is not among the ETF tax strategies they currently view as abusive. But it lays bare a widening gap between a US tax code written for a market in which domestic households were the dominant owners, and a market in which non-US investors now hold more than 30 percent of publicly traded US stocks.
The Mechanics: How a Dividend Becomes a Capital Gain
The trade exploits the way index funds handle dividends. The 500 companies in the S&P 500 pay cash dividends on scattered dates through the year; index funds collect that cash and pay it out to shareholders once a quarter. The index's annual dividend yield is just over 1 percent of its market value — thin by historical standards, but on tens of billions of dollars of notional, and after a 30 percent haircut at the border, worth moving.
To avoid receiving a fund's dividend, an investor must have sold the position by market close on the day before the ex-dividend date — the point at which the ETF's shares begin trading without the right to the upcoming payment. When a fund goes ex-dividend, its net asset value drops by roughly the amount of the distribution. An investor who holds the fund that has not yet gone ex owns an asset whose price still carries the accrued dividend; an investor who switches into the fund that has already gone ex buys at a post-payout discount. The total return is unchanged, but its character has shifted: what would have been a dividend, taxable at 30 percent, arrives instead as price appreciation. For investors who do not reside in the United States, capital gains on securities trades are generally not subject to US tax.
"S&P 500 ETFs are used for sophisticated trading," said Matt Bartolini, global head of research strategists at State Street Investment Management. The flow patterns show some large institutions "are utilizing the ETFs to gain continuous exposure without taking receipt of the dividend," he said.
The switching shows up in both fund flows and trading volumes, which spike around IVV's ex-dividend date each quarter. A mirror-image pattern appeared for the first time in June around State Street's SPYM, the SPDR Portfolio S&P 500 ETF, after State Street changed the timing of that fund's dividend payouts in a way that made toggling with the flagship SPDR S&P 500 ETF Trust easier. Bartolini said SPYM's growth has made it "more of an institutional tool." BlackRock declined to comment on why it altered IVV's dividend timing.
The phenomenon is not confined to equities. BlackRock's iShares 0-3 Month Treasury Bond ETF, which distributes interest monthly, sees as much as $1.3 billion pulled just before its ex-dividend date, with a similar amount flowing back a day later. The common thread is not asset class but tax status: any distribution that would be taxable to a non-US holder creates an incentive to step aside for a day.
Why This Is Structural, Not a Passing Quirk
The first question the pattern raises is whether it is a cyclical misalignment that market plumbing will eventually smooth away, or a structural feature of the modern ETF market. The answer is structural, for three reasons.
First, the tax itself is a durable feature of US law. Dividends paid to nonresident aliens are fixed, determinable, annual or periodical income, taxed at a flat 30 percent at source unless a treaty lowers the rate, with no deductions allowed against the income. That is not a temporary provision; it is the standing cost of owning US dividend-paying assets as a foreign person.
Second, the arbitrage depends on market structure that is not going away. The S&P 500 is now tracked by several enormous, deeply liquid funds that are economically interchangeable but operationally independent — each issuer sets its own distribution calendar. Competition on fees and liquidity, good for investors, produced the calendar misalignment that makes the switch possible. There is no central scheduler for ETF dividends, and no incentive for any single issuer to move first.
Third, the ownership base that makes the trade worthwhile has shifted permanently. More than 30 percent of US publicly traded stocks are now owned by non-US investors, according to a 2024 analysis. As long as foreign ownership stays high and the withholding rate stays at 30 percent, the incentive to avoid the levy is permanent. The quarterly flow may ebb and swell with the size of distributions and fund assets — that is the cyclical layer riding on top — but the mechanism will not self-correct.
The Second-Order Consequence: Issuer Competition Built the Loophole
The obvious reading of this story is that foreign investors are clever and the tax code is leaky. The less obvious reading is that the ETF industry's own success constructed the arbitrage. Issuers competed to offer the cheapest, most liquid S&P 500 fund, and in doing so fragmented one index into rival products with independent operational calendars. The tax dodge exists precisely because the products are close enough substitutes to swap without cost, and different enough in their mechanics to create a gap to step through.
That creates an awkward alignment of incentives. Nothing in any issuer's commercial interest argues for harmonizing dividend calendars — doing so would close a feature some of their largest institutional clients value. And because the switch preserves continuous market exposure, it does not distort the price of the underlying index in any lasting way. It is a transfer of tax liability, not a bet on direction.
The second-order cross-market implication runs through Treasury's own prioritization. Officials have been scrutinizing a menu of ETF tax strategies, including funds that offer a pre-packaged version of the rotation using in-kind mechanics to avoid capital gains as they move between similar holdings. But when asked at a July industry conference whether their concern extended to foreign investors rotating between similar investments directly, one official, Erika Nijenhuis, answered plainly: "That's not a focus."
That distinction — packaged product versus direct trade — is the line the market will now test. After the conference, both Jeffrey Hochberg, a Sullivan & Cromwell tax lawyer who posed the question, and Steven Rosenthal, former counsel to the Joint Committee on Taxation, said separately it was the right call.
"It just seems like you're selling an S&P for another S&P," Rosenthal said in an interview. "I would treat this as just fine from a tax perspective."
The Counter-Thesis: It Is Just Selling One S&P for Another
The strongest argument against treating the shuffle as a problem is also the simplest: the investor's economic position never changes. They are long the S&P 500 before, during and after the trade; they take the same total return; they simply receive it as price appreciation instead of as a distribution. Under that view there is no avoidance of economic substance, only a choice between two forms of the same return, and the US tax base is not truly eroded because the underlying companies paid their taxes regardless.
That argument carries real weight, and it explains why Treasury has not made direct rotation a target. But it understates what the withholding tax is meant to do. The 30 percent levy is a backstop on income flowing to owners outside the US tax net; when that income is systematically converted into untaxed capital gains, the backstop fails for exactly the investor class it was designed to reach. The revenue at stake today — on the order of $147 million a year for the IVV-VOO pair alone — is small relative to the federal budget, but it scales with ETF assets and with foreign ownership. A practice that is "just fine" at $150 million a year is harder to dismiss if it grows several-fold as the ETF market expands.
The history of dividend avoidance also argues against complacency. A generation ago, banks sold loan and swap products to hedge funds to minimize the same withholding tax. Congress responded to a 2008 Senate investigation with new anti-abuse provisions, and the tax authority later won settlements from funds that used those structures. The difference now is that the arbitrage needs no engineered product — only two funds that already exist and a calendar to read.
What Would Prove This Wrong
Two observable signals would falsify the structural call. First, if IVV and VOO were to align their ex-dividend calendars, the arbitrage window would close and the quarterly flows would vanish — evidence that the pattern was plumbing, not regime. Second, if Treasury or the IRS were to issue guidance treating economically identical rotations around ex-dividend dates as dividend-stripping-type transactions, the tax benefit would disappear and with it the incentive. A third, softer signal: if the estimated annual tax saving grew well beyond its current level — say past $500 million — without a regulatory response, the political tolerance for the practice would be higher than this analysis assumes.
What to Watch: Three Time Horizons
Short term (next quarter): the middle of September should bring the next predictable IVV-to-VOO flow episode, followed by a reversal. Trading volumes in both funds should spike around BlackRock's ex-dividend date. Any deviation from the pattern — a smaller flow, or no reversal — would be the first sign the trade is losing its edge.
Medium term (6 to 18 months): watch whether State Street's SPYM continues to gain assets as an institutional tool and whether the SPYM-SPY toggle repeats around its distribution dates. Also watch Treasury's list of "potentially abusive" ETF strategies for any expansion beyond packaged products to direct rotations.
Long term (structural): the durability of the practice rests on three pillars — the 30 percent withholding rate, foreign ownership above 30 percent of US equities, and misaligned distribution calendars across interchangeable mega-funds. Remove any one and the trade dies. Keep all three and it becomes a permanent feature of how the world owns US stocks.
The base case is continuity: the shuffle repeats each quarter, Treasury keeps its hands off direct rotation, and the practice grows quietly with ETF assets. The upside case for the tax authority is a guidance change or voluntary calendar harmonization that closes the gap. The downside case is escalation — a politically charged moment in which a nine-figure annual leakage becomes a symbol of a tax code that foreign investors can read more profitably than domestic ones.
The real story here is not that investors found a clever way around a tax. It is that the world's largest equity market now runs on infrastructure built for a different era of ownership, and the gap between the two is measured in tens of billions of dollars moving on schedule, four times a year, with nobody trying very hard to stop it.
Explore more exclusive insights at nextfin.ai.
