A founder can die holding $2 billion in appreciated stock and $600 million in loans against it, and the income tax owed on that gain comes to zero. Larry Ellison has run this play against his Oracle shares for years: borrow against the stock instead of selling it, spend the cash tax free, then let heirs inherit at a stepped up basis that erases the gain entirely. A wage earner's 401k withdrawal gets taxed every year with no equivalent escape hatch. That's the asymmetry the Budget Lab at Yale has spent two years trying to model a fix for. The real question is whether the IRS can close a mechanism built from three provisions that each look defensible alone.
The strategy has a name now: buy, borrow, die. Buy an appreciating asset, borrow against it tax free while you're alive, then let your heirs inherit it at a stepped up basis when you die, wiping out the unrealized gain for tax purposes entirely. The Budget Lab at Yale, which published a 2025 study modeling reform options for this exact mechanism, frames it plainly: this isn't a loophole in the sense of a drafting error. It's the predictable output of three separate tax provisions interacting exactly as written.
Multi billionaire founder executives holding low basis, highly liquid shares in a handful of the world's most valuable companies benefit most. Nobody else has both the asset concentration and the lending relationships to make this work at scale. Here's how each of those three provisions works on its own.
Why Does Borrowing Against Stock Beat Selling It?
Selling vs Borrowing: Tax Outcome on 50 Million Dollars
| Action | Amount Accessed | Federal Tax Rate | Tax Owed |
|---|---|---|---|
| Sell Stock | $50,000,000 | 23.8% | up to $11,900,000 |
| Borrow Against Stock | $50,000,000 | 0% | $0 |
| Per Share Example ($1 to $100) | $99 gain / share | Sell: taxed / Borrow: not taxed | Borrow wins |
Source: Article estimates based on federal capital gains rate and securities based lending mechanics
Sell $50 million of Tesla stock and the tax bill runs up to 23.8 percent in federal capital gains tax, plus whatever the state charges, on any appreciation since purchase. Borrow $50 million against that same stock from a private bank's securities based lending desk, and the IRS gets nothing. Loan proceeds are not income under the tax code. They never have been. That single asymmetry is the entire mechanism.
The math gets sharper the longer the position is held. Suppose a founder bought shares at $1 and they are now worth $100. Selling triggers tax on $99 of gain per share. Borrowing against the full $100 value triggers tax on zero. The lender doesn't care about the tax basis, only the collateral value and the loan to value ratio. Private banks offering wealth management clients securities based lending have generally provided loan-to-value ratios that vary widely, depending on concentration and volatility of the underlying portfolio.
Interest rates on these loans have moved with the broader rate environment. When the Fed held rates near zero through 2021, securities based lines were priced attractively enough that some borrowers effectively paid less in interest than they would have paid in capital gains tax on an equivalent sale. Even in 2026, with policy rates elevated compared to the previous decade, the calculus often still favors borrowing when the alternative is a large, one time capital gains bill combined with the loss of future appreciation on the sold shares.
The Budget Lab treats this as a portfolio distortion question as much as a revenue question. Households that would otherwise diversify out of a concentrated stock position instead hold on, because selling costs tax and borrowing doesn't. The tax code is quietly pushing wealthy shareholders toward concentration risk, not away from it. An unrealized gain that's never realized gets taxed at zero, every time. Borrowing solves the tax problem while the founder is alive. What happens to that loan when the founder dies is the second provision in the mechanism.
What Happens To The Loan When The Borrower Dies?
The Buy, Borrow, Die Mechanism
Acquire appreciating stock at low basis (e.g. $1 per share)
Take securities based loans against stock value. Loan proceeds are not taxable income.
Live on borrowed cash tax free instead of selling shares
Heirs inherit at stepped up basis. Unrealized gain is erased for tax purposes.
Source: Budget Lab at Yale, 2025 study on tax reform options
Consider a founder who dies still holding $2 billion in appreciated stock and $600 million in loans secured against it. Under current law, the heirs receive the stock at its fair market value on the date of death. That stepped up basis erases the original unrealized gain for capital gains purposes. The estate may owe estate tax on the net value above the exemption threshold, which sits well above $13 million per individual as of 2026 and is indexed for inflation. But the income tax bill on decades of appreciation simply disappears.
This is the die step, and it's where the strategy converts a deferred tax liability into a tax liability that never materializes. Compare that to an ordinary worker whose 401k withdrawals are taxed as ordinary income every year, with no equivalent mechanism to step up basis and skip the bill. The Budget Lab's report calls this out directly as horizontal inequity: two households with the same economic income, one from wages, one from unrealized capital gains funded through borrowing, face wildly different effective tax rates.
Estate tax is supposed to be the backstop here, and for many estates it functions that way. But founder held stock in fast growing companies often keeps appreciating faster than the estate tax exemption grows, and sophisticated estate planning, including grantor retained annuity trusts and dynasty trusts, has historically been used to move appreciation outside the taxable estate altogether. Some legal scholars describe this less as a single loophole and more as a stack of provisions that were each defensible in isolation but combine into something nobody voted for as a package.
The first two provisions, tax free borrowing and the stepped up basis at death, are settled law with predictable outcomes. The third piece of the mechanism is less settled: whether the IRS has any tool that reaches unrealized gains before death at all.
Can The IRS Actually Tax Unrealized Gains?
The Founder's Estate at Death: A Case Example
Appreciated Stock Held
$2.0B
Loans Secured Against It
$600M
Income Tax Owed on the Gain
$0
Stepped up basis at death erases the unrealized gain entirely
Source: Article illustrative example based on Budget Lab at Yale mechanism analysis
Take a hypothetical household holding $200 million in a single publicly traded stock. A wealth tax approach would require the IRS to assign an annual value to that position, which is straightforward for liquid public equity but gets genuinely hard once family businesses, private equity stakes, real estate, and closely held companies enter the picture. The Budget Lab's report treats this as the central design tension: progressivity versus administrability.
A pure net worth test would, by construction, hit the households buy borrow die is aimed at. It would also require the IRS to value complex, illiquid portfolios every year, an agency already stretched thin on enforcement resources relative to the complexity of high end tax planning. Wealthy taxpayers have historically responded to valuation based tax rules by restructuring ownership, layering entities, and shifting assets into vehicles designed to complicate valuation. A wealth test built for progressivity can end up creating a new market for avoidance built around exactly that difficulty.
The alternative the report weighs is narrower: tax the loan itself, or tax borrowing against appreciated assets above a certain threshold as a realization event, rather than trying to value the entire estate every year. Some proposals under discussion would trigger capital gains recognition once loans secured by appreciated securities cross a set dollar threshold, or would tax the interest deduction differently when the loan isn't used for a business purpose. Each version tries to hit the mechanism directly rather than net worth broadly.
None of these options come free of tradeoffs. A threshold based rule creates a new cliff for tax planners to engineer around, structuring multiple smaller loans instead of one large one. A rule aimed only at the largest, most liquid public company holdings misses privately held wealth entirely. Administrability, not progressivity, usually ends up being the binding constraint on how aggressive a reform can get, and there's no clean number attached to that tradeoff. Any rule built to hit billionaire scale borrowing has to pass through the same lending mechanism that ordinary investors use at a much smaller scale, which is exactly where the political difficulty starts.
Who Actually Pays For This Tax Reform Debate?
When Borrowing Beats Selling: Interest Rate vs Capital Gains Tax
| Environment | Loan Interest Cost | Advantage vs Selling |
|---|---|---|
| 2021, Near Zero Fed Rates | Very Low | Strongly Favors Borrowing |
| 2026, Elevated Policy Rates | Moderate | Still Favors Borrowing |
| Selling and Paying Tax | 23.8% Tax Due | Loses Future Appreciation Too |
Source: Article analysis of securities based lending economics across rate environments
Look at Meta's early employees who took loans against restricted stock units before a lockup expired, a modest, low stakes version of the same mechanic used at billionaire scale. That comparison matters because it shows buy borrow die isn't exotic. It runs on a spectrum, from ordinary securities based lines of credit that millions of upper middle class households use to fund a home purchase, up through billion dollar credit facilities collateralized by founder stakes in companies they control.
That spectrum is exactly why reform is politically harder than it looks on paper. A rule written broadly enough to hit the multi billionaire founder executive case will also touch smaller investors using securities based lending for ordinary purposes, unless the design carves out thresholds carefully. A rule written narrowly enough to avoid that problem risks becoming easy to route around, since sophisticated tax planning has repeatedly shown an ability to restructure around thresholds within a filing cycle or two.
Congress has floated versions of billionaire minimum tax proposals in recent years, including mark to market elements aimed at tradable assets held by ultra high net worth households. None have passed into law as of September 2026. Each has run into the same three objections: liquidity for taxpayers who would owe tax on paper gains, valuation complexity for illiquid holdings, and constitutional questions about whether unrealized gains constitute income under the Sixteenth Amendment, a question the Supreme Court has touched on but not resolved definitively.
Every year buy borrow die remains untouched, the gap between how wage income and unrealized capital gains are taxed widens further for the households best positioned to exploit it. The founder who dies holding $2 billion in stock and $600 million in loans against it, the case this piece opened with, still owes zero income tax on that gain under every rule Congress has passed as of September 2026. Zero laws passed, zero tax owed. That's the actual answer to whether the IRS can close this mechanism: not yet, and not by any method currently on the table.
This article is for informational and educational purposes only and does not constitute financial, investment, or legal advice. The views expressed are analytical observations and should not be relied upon for personal financial decisions. Always consult a qualified financial advisor before making investment decisions.