Buy, Borrow, Die: How a Lifetime of Gains Goes Untaxed, and How to Stop It

Every dollar you earn at work gets taxed before it hits your bank account. Federal income tax, state tax, Social Security, Medicare. You never even see the money.

Meanwhile, the wealthiest Americans have a legal way to spend millions of dollars a year and owe no income tax on any of it. Not a reduced rate. Zero.

It's called Buy, Borrow, Die. And it's not a conspiracy theory. Financial advisors now explain it openly on YouTube, complete with QR codes to book a strategy session. The loophole has gone retail.

We should close it. Here's how it works, what it costs the rest of us, and what the fix looks like.

How Buy, Borrow, Die Works: The Three-Step Machine

The strategy runs on three ordinary tax rules that combine into something extraordinary.

Step one: Buy. Purchase assets that appreciate. Stocks, businesses, real estate. Under current law, unrealized gains are not taxed. Your Nvidia position can grow from $25,000 to $250,000 and you owe nothing until you sell.

Step two: Borrow. When you need cash to live on, don't sell. Borrow against the portfolio instead. Loans are not income, so borrowed money is never taxed. You get cash to spend, your investments keep compounding, and in some cases you can even deduct the interest.

Step three: Die. When you pass, your heirs inherit the assets with a stepped-up cost basis. The basis resets to the market value on the date of your death. That $225,000 gain? Erased. Your estate repays the loans from the untaxed assets, and your heirs can sell the next day tax-free.

Nobody pays capital gains tax on the appreciation. Not you during your life. Not your heirs after. The gain simply vanishes from the tax system.

The estate tax is supposed to be the backstop here, and it is worth being straight about it: estates above $15 million per person do pay 40 percent. But it reaches almost no one. In the most recent year of complete IRS data, about 2,100 estates paid it out of roughly 2.8 million deaths, fewer than one in a thousand. And it is a tax on the estate, not on the gain. The capital gain itself is still never taxed, by anyone, ever.

USC law professor Edward McCaffery coined the term decades ago. ProPublica documented the results in 2021: the 25 wealthiest Americans grew their fortunes by $401 billion between 2014 and 2018 while paying $13.6 billion in federal income tax. A true tax rate of 3.4 percent. Jeff Bezos paid zero federal income tax in 2007 and 2011. Elon Musk paid zero in 2018.

That data stops in 2018 for one reason: it came from a leak, not from anything the public is allowed to see. No one has released comparable individual returns since. What we do have points the same direction. A White House analysis using a similar approach put the 400 wealthiest families' effective federal income tax rate at 8.2 percent once the growth in their stock is counted as income.

Be precise about what that 3.4 percent measures, because critics are right that it is not a conventional tax rate. It is tax paid divided by how much richer they got. Measured the ordinary way, against the income they actually reported, those same 25 people paid about 15.8 percent. A teacher earning $60,000 pays about 16 percent in federal income and payroll taxes, so on that measure they are nearly even.

Do not skip past that. The gap between 3.4 and 15.8 is not a gotcha against the argument. It is the argument. One ratio is measured against everything they gained; the other only against the sliver of it the code ever counted. The whole machine described above exists to keep that sliver small.

So here is the comparison that needs no special math. Two people each spend $2 million this year. The first sells stock to do it and pays about 23.8 percent on the gain. The second borrows against the same stock and pays nothing at all. The borrower still owes the money and pays interest on it, so this is not free. But only one of them owes tax, and only one of them had to give up the asset.

And the direction of travel looks one way. Emmanuel Saez and Gabriel Zucman estimate that the 400 richest American families paid an overall rate around 56 percent in 1960, 47 percent by 1980, and 23 percent in 2018, which by their measure is the first time in a century the top 400 paid a lower rate than the bottom half of the country.

Treat that series as contested rather than settled. It is the most argued-over statistic in American tax economics. Gerald Auten and David Splinter, using different assumptions about unreported income and refundable credits, get materially higher numbers for the top and a much flatter trend. The direction is widely accepted; the magnitude is not. Buy, Borrow, Die is one of the mechanisms behind it either way.

Is Buy, Borrow, Die Legal? Yes, and That Is the Problem

Defenders say nothing here is illegal. Correct. That's exactly the point.

The problem is structural. Our tax code makes one foundational assumption: income means realized income. That assumption made sense when it was hard to spend wealth without selling it. It stopped making sense the moment banks started handing out portfolio-backed credit at scale.

For wage earners, income and spending power are the same thing, and both get taxed. For the wealthy, the tax code lets them separate the two. They get the spending power of their gains through loans while the taxable event never arrives, and then death deletes it entirely.

A loan against an appreciated portfolio is functionally a sale. You've converted your gains into spendable cash. The only difference is the paperwork. We tax the substance of every other transaction. We should tax the substance of this one.

What the Step-Up in Basis Costs: $72 Billion a Year

Start with the official number. The Joint Committee on Taxation scores the exclusion of capital gains at death, the "die" step, at $72.5 billion in forgone revenue for 2026. Across 2025 to 2029 it comes to $379 billion, which puts it among the ten largest individual tax breaks in the entire code.

For scale: the federal government spends $18.4 billion a year on Title I, the entire program supporting low-income K-12 schools nationwide. The tax erased at death each year is worth nearly four Title I programs. Put in the terms this site was built for, $72.5 billion a year is roughly 980,000 teacher salaries at the national average of $74,000.

But the official number only counts the death step. The full machine costs more, because the "buy" and "borrow" steps let gains compound untaxed for decades before death erases them.

Here's the fuller math:

The stock of untaxed gains. Americans for Tax Fairness, analyzing Federal Reserve data, found that billionaires and centi-millionaires held at least $8.5 trillion in unrealized capital gains as of 2022. That's 64,000 households, about one in every 2,000 in the country, holding 18 percent of the nation's unrealized gains. At the top capital gains rate of 23.8 percent, that's roughly $2 trillion in tax already embedded in existing fortunes. Buy, Borrow, Die is the machine built to ensure it never gets paid.

The annual flow. Billionaire wealth alone has more than tripled since 2017, from $2.9 trillion to over $9 trillion, averaging roughly $650 billion in new gains per year. Tax that flow at 23.8 percent and you get about $150 billion a year accruing to billionaires alone that the current system will never collect, before counting centi-millionaires.

No realistic reform captures all of that. Exemptions, valuation rules, and behavioral responses shrink every estimate, which is why Treasury scored its capital-income package, taxing gains at ordinary rates for high earners and treating death as a realization event, at a combined $322 billion over ten years. But the honest range is clear:

Those are measures of what goes untaxed, not of what any bill would collect. Scored reforms are smaller, and the article returns to that below. But whatever is not collected here is collected somewhere else, or borrowed.

The Banks Advertise This

For years, the practical brake on Buy, Borrow, Die was borrowing cost. At the big retail brokers a margin loan runs close to 12 percent on a small balance today. Fidelity's published schedule falls to 7.5 percent once you are borrowing over a million; Schwab stops publishing above $500,000 and tells you to call. At those rates, borrowing can cost more than the tax it avoids.

The wealthy solved that too. The current tool of choice is a box spread: a bundle of options contracts that functions exactly like a fixed-rate loan, priced off institutional options markets at a fraction of a percent above Treasury yields. In early August 2026 the implied rate ran about 4.2 to 4.7 percent across terms from one month to five years, a quarter to a half point above Treasuries. Against Schwab's or Fidelity's published rate on a $100,000 balance, that is less than half the cost. One honest qualifier: Interactive Brokers already charges about 5 percent on that same balance, so against the cheapest broker in the market the saving is a fraction of a point, not half. The gap this exploits is less options wizardry than the markup the household-name brokers charge, and the fact that the largest portfolios know how to step around it. Tony Yang, now a co-founder of SyntheticFi, a firm that sells the trade to financial advisors, told Bloomberg he found the strategy on Reddit and built it himself in his Schwab account in late 2021, borrowing about $650,000 at 1.6 percent, locked for five years, toward a Bay Area house without selling a share. Rates were near zero then and that number is not repeatable today. The mechanism is.

The banks are not hiding this. They advertise it. Merrill tells clients its Loan Management Account can "help you avoid the tax consequences of selling appreciated investments." Schwab's promotional video for its Pledged Asset Line lists "Avoid tax consequences" on screen. J.P. Morgan's private bank tells clients to "avoid selling so your allocation stays on target and your high-conviction positions can keep compounding." Its wealth management arm discloses separately that "Chase Bank and its affiliates may earn more if you borrow against your securities and other assets rather than liquidate assets."

The volume backs up the pitch. Merrill clients took on $10 billion in new loans during 2025 alone, and the balances at every major firm are climbing fast.

The rates fall as the loan grows, the way secured credit usually does. Schwab charges its smallest borrowers about four percentage points over the benchmark and its largest about two. What is unusual is what that discount is buying: the bigger the position, the cheaper it gets to avoid ever paying tax on it.

You can see it in the filings. Oracle disclosed in September 2025 that Larry Ellison had pledged 346 million Oracle shares as collateral "to secure certain personal indebtedness." Meta bars its people from pledging company stock for loans unless the board's governance committee approves it. The board has approved exactly one such framework, for Mark Zuckerberg, who has 12 million Class B shares pledged. When a company makes everyone ask permission and then grants it to exactly one person, it is telling you what the practice is worth.

Watch what financial advisors pitch alongside it. Borrowing instead of selling keeps your taxable income near zero on paper, which can qualify wealthy retirees for ACA premium subsidies and open cheap Roth conversion windows.

The rules here just changed, and it cuts both ways. The enhanced premium tax credits expired on December 31, 2025, and the Senate effort to revive them collapsed in February 2026. The 400 percent federal poverty level cliff is back, so a household now needs reported income under roughly $62,600 for one person, or $128,600 for a family of four, to get any help at all. That kills the most extreme version of this, where someone reporting well into six figures still collected a subsidy.

But it sharpens what remains, because a cliff is a cliff. One dollar of realized gain over that line now costs a household its entire subsidy, worth hundreds a month and more for older enrollees. Someone spending from a loan decides exactly which side of the line they land on. Someone living on a paycheck has no such dial. Assistance aimed at households that cannot afford coverage stays available to people with seven-figure portfolios, because the code measures what showed up on the return rather than what the household can actually spend. Not by cheating. By following the rules as written.

Now the honest caveat, because the best available research cuts against the emphasis of this section. Edward Fox and Zachary Liscow, writing in the Journal of Public Economics in 2025, looked at how much the very rich actually borrow and found it is small: new borrowing runs about 1 to 2 percent of the economic income of the top 0.1 percent. Their conclusion is that the dominant strategy is closer to "buy, save, die." The ultra-wealthy mostly do not need to borrow, because their taxed income already covers anything they care to spend.

That finding does not rescue the loophole. It relocates it. If they are not borrowing much, they are simply holding, and the gains still sit there untaxed until death erases them. It means the borrow step is the smallest of the three, and the die step is the one that matters. Which is exactly how the fixes below are ordered.

The Other Side of the Ledger

Be careful with that $72.5 billion, because it is a tax expenditure estimate, the gain the code declines to count, not a revenue score. Actual scores of actual bills build in exemptions and behavior and come in lower, roughly $11 to $54 billion a year depending on design.

The high end is close to three Title I programs. Here is who that matters to.

37 percent of American adults could not cover a $400 emergency expense with cash, per the Federal Reserve. 100 million Americans carry $220 billion in medical debt, 41 percent of all adults. 43 percent of adults skipped, rationed, or substituted prescription medication in the past year because of cost.

Hold those numbers next to each other. Unpaid medical bills in this country come to roughly $220 billion. The tax sitting uncollected inside the top 0.05 percent's untaxed gains is roughly nine times that.

Meanwhile the fortunes on the other side keep compounding. As of June 2026, 977 American billionaires hold $9.24 trillion, up $2.2 trillion in a single year, and the 16 richest of them now hold more than every U.S. billionaire combined held in September 2020. At that scale money stops being spending power and becomes something else: compounding collateral, and the political influence to keep it compounding.

This is not a story about envy. It is a story about a tax code that taxes work in full, taxes extreme wealth almost not at all, and then asks the people cutting their pills in half to cover the difference.

How to Close the Buy, Borrow, Die Loophole

Three fixes. The first is the one that matters most, for the reason the last section gave: if the wealthy mostly hold rather than borrow, the step-up is what makes holding permanent. The first is also ordinary tax plumbing that other countries already use. The second has no international precedent that I could find, and the third is the most legally exposed.

One honest note before them. A step-up in basis at death is not some American peculiarity: it is the most common approach in the OECD, used by the UK, France, Ireland and others. What is unusual here is the combination. Those countries pair a step-up with an inheritance tax that actually reaches ordinary estates, at thresholds in the tens or hundreds of thousands. The United States pairs an unconditional step-up with a $15 million exemption, so the relief applies to every estate including the 99.9 percent that owe no estate tax at all. The OECD's own recommendation is to reconsider the step-up precisely "where inheritance or estate tax exemption thresholds are very high."

First, end the stepped-up basis for large estates. This is the linchpin. Treat death as a realization event: unrealized gains above a generous exemption get taxed before assets pass to heirs. Canada has done exactly this since 1972, and it works administratively, which was the thing critics said was impossible. Be honest about the rest of the Canadian record, though: Ottawa repealed its estate tax the same year, and the new tax on gains at death raised far less than the old estate tax had. Taxing the gain is not a substitute for taxing the estate. It is the thing you do so that a lifetime of appreciation does not vanish entirely.

This is not hypothetical. It was in Biden's budget, and it is sitting in Congress right now as the Equal Tax Act, introduced in September 2025, with a $1 million exemption, an extra carve-out for family farms and businesses, and five years to pay. A nearly identical provision was stripped out of Build Back Better in 2021 after roughly 330 agricultural groups lobbied against it.

Here is what that lobbying was protecting people from. The Tax Policy Center estimated that under a $1 million exemption, fewer than 3 in 100 households headed by someone over 70 would owe anything at all. The USDA ran its own numbers on the family farm question and found that about 1 in 100 farm estates would owe capital gains tax at death, while more than 80 percent would owe nothing and keep the full step-up.

Second, treat large asset-backed borrowing as realization. When someone borrows against appreciated securities, tax the embedded gain on the collateral at that point, with basis adjusted so nothing gets taxed twice. You want to borrow $650,000 against your stock to buy a house? Fine. You've just realized $650,000 of your gains, same as if you'd sold. Written this way it reaches the "borrow" step without touching an ordinary mortgage or a small business line of credit, though as the next paragraph shows, the bill actually before Congress is drawn more broadly than that.

Senator Ruben Gallego introduced a version of this in June 2026, reaching only people who clear $100 million in income or a billion in assets in each of three straight years. His bill is broader than the description above: it treats any large loan to those taxpayers as a realization event, not only one secured by appreciated stock. The Yale Budget Lab scored a version of it at $102 billion over ten years, and found that borrowing currently carries about a 12 percentage point tax advantage over selling. Yale is also honest about the ceiling: borrowing is a small share of what the very top actually live on, so this fix closes a door rather than filling the treasury.

Third, a minimum tax on the very top as a backstop. For households above $100 million, an annual minimum tax on total income including unrealized gains, creditable against future capital gains taxes so it functions as a prepayment, not a double tax. This is the Billionaire Minimum Income Tax framework, and it exists precisely because the top 0.001 percent can wait out any realization-based rule.

Notice what these have in common. Buy, Borrow, Die is a chain, and a chain only works if every link holds.

The standard objections have standard answers. Valuation is hard for private assets? Then apply the rules to publicly traded securities first, where the price updates every second. Liquidity concerns? Payment plans, same as we already offer estates. It punishes success? A capital gains tax on gains you actually collected and spent is not punishment. It's the same deal everyone else already lives under.

There is one serious objection, and it is about the Constitution. In Moore v. United States, decided in 2024, the Supreme Court upheld the tax in front of it but pointedly refused to decide whether income must be realized to be taxed. Four justices went out of their way to say it must be. That is one vote short of a majority that could strike down any annual tax on unrealized gains.

Which is precisely why the order of these fixes matters. Ending the step-up and taxing large borrowing are both triggered by an event, a death or a loan, exactly like every other realization rule in the code. They sit on the firmest ground available. The minimum tax is the ambitious one, and it is the one most likely to end up in front of the Court.

The Choice We're Making

Buy, Borrow, Die isn't a glitch someone forgot to patch. It survives because the people who benefit from it fund the politics that protect it. Every year Congress doesn't act is a choice to keep taxing work at up to 37 percent while taxing this at zero.

It survives despite the public, not because of it. When Fox News asked in September 2025 whether the government should raise taxes on the wealthy to strengthen social programs, 73 percent said yes, including 56 percent of Republicans. Gallup finds 58 percent say upper-income people pay too little. Pew finds the top two frustrations Americans have with the tax system are that wealthy people (61 percent) and corporations (60 percent) don't pay their fair share.

What almost nobody polls on is the machinery. Ask Americans about stepped-up basis, or about treating a giant loan as the sale it functionally is, and you will find almost no public data at all. The principle is settled and has been for years. The plumbing is where the lobbying happens, in language technical enough that most people never see the decision being made.

Meanwhile, the estate tax now exempts $15 million per person, $30 million per couple as of 2026. The last backstop keeps getting weaker while the avoidance machine keeps getting cheaper.

Tax systems run on one thing: the belief that everyone is playing the same game. Right now, provably, we are not. A wage earner pays tax on every dollar before seeing it. A billionaire spends millions a year and dies having never paid tax on the fortune that funded it.

Close the step-up. Tax large borrowing as the sale it really is. Backstop the top with a minimum tax.

Same game. Same rules. That's the whole ask.