Taxing Money That Doesn't Exist
On February 12, the Dutch House of Representatives passed a law that taxes you on money you haven't made. The Wet werkelijk rendement box 3 β the Actual Return Act β imposes a 36% annual tax on the unrealized appreciation of listed stocks, bonds, and cryptocurrencies. If your portfolio rises by EUR 10,000 on paper, you owe EUR 3,600 in tax. You didn't sell anything. You didn't receive a dividend. You have no cash. But you owe the government 36% of a number on a screen.
Thirteen days later, the new Dutch government announced the bill "cannot pass as is."
The speed of the reversal tells you something. But the more interesting question is why anyone thought this would work in the first place β and what the attempt reveals about a conceptual problem that tax systems everywhere are struggling with.
The Ghost in the Portfolio
The core issue is deceptively simple: what counts as income?
If you buy a stock for EUR 100 and it rises to EUR 150, you are EUR 50 richer β on paper. You could sell and pocket the gain. But you haven't. The gain is unrealized. It exists as a number in a brokerage account, contingent on market conditions that could reverse tomorrow. Taxing it requires treating a fluctuating estimate as a concrete fact.
The Netherlands isn't the first to try. Norway taxes net wealth at 1.1% annually. The Biden administration proposed a 25% minimum tax on unrealized gains for households worth over $100 million. Senator Wyden's Billionaires Income Tax Act would end the "buy, borrow, die" strategy by taxing paper gains annually. None of these passed in the US. The Dutch bill passed β and then immediately started unraveling.
The reason is a brutal asymmetry built into any mark-to-market tax system.
The Asymmetry Problem
Consider a simple scenario. You hold EUR 100,000 in stocks. In year one, the market rises 30%. Your portfolio is now worth EUR 130,000. Under the Dutch system, you owe 36% on the EUR 30,000 gain: EUR 10,800. You sell some shares to pay the tax.
In year two, the market drops 25%. Your portfolio falls back to roughly where it started. You've lost your gains. But you've already paid EUR 10,800 in tax on gains that no longer exist.
The Dutch law allows unlimited loss carry-forward β you can offset future gains against past losses. But there is no loss carry-back. You cannot reclaim taxes paid on gains that subsequently evaporated. Over a volatile decade, an investor can end up paying substantial tax on returns that, cumulatively, were modest or negative. The tax captures the peaks but doesn't refund the valleys.
This isn't a theoretical concern. Anyone who held tech stocks through 2021-2022 watched portfolios double and then halve. Under mark-to-market taxation, you'd have paid tax at the top and received nothing at the bottom. The carry-forward helps eventually β if the market recovers, if you live long enough, if you don't need the liquidity for something else. But it doesn't solve the fundamental problem: the government captures a share of every up-move in real time while deferring its share of every down-move indefinitely.
The Liquidity Trap
There's a second problem, more practical but equally corrosive. Unrealized gains are, by definition, illiquid. To pay a tax on paper profits, you must either sell assets (crystallizing the very event taxation was meant to defer) or find cash elsewhere. For a diversified portfolio of listed stocks, this is inconvenient but feasible. For less liquid assets β startup equity, real estate, private funds β it becomes a genuine crisis.
The Dutch law partly addresses this by putting real estate and qualifying startup shares on a realization-only track. But the qualifying criteria are narrow: the company must be less than five years old, revenue below EUR 30 million, and the investor must hold less than 5%. A six-year-old biotech startup with EUR 35 million in revenue doesn't qualify. An angel investor with a 7% stake doesn't qualify. The exemptions acknowledge the problem without solving it.
The Innovation Problem
Capital flight gets the headlines, but there's a subtler issue that matters more in the long run. The modern economy runs on unrealized gains. Not as a tax loophole β as a structural feature.
A software company's value is almost entirely intangible: code, network effects, data, brand. These assets don't generate proportional cash flow during their growth phase. A biotech firm spends a decade burning cash before a single drug reaches market. A deep-tech startup might not produce revenue for years while its patents and prototypes appreciate in value. The gap between paper wealth and available cash isn't an edge case. It's the default state of innovation.
Taxing unrealized gains penalizes exactly the behavior that produces breakthroughs: concentrated positions in illiquid, high-variance ventures held over long time horizons. An angel investor who puts EUR 100,000 into a startup valued at EUR 1 million watches that company get revalued at EUR 10 million in a funding round three years later. Under mark-to-market taxation, the investor owes 36% on EUR 900,000 β EUR 324,000 β despite having received nothing and having no ability to sell. The only options are to sell the stake (usually impossible before an exit), find the cash elsewhere, or never make the investment in the first place. Most will choose the third option.
The Dutch law's startup exemptions β company under five years old, revenue under EUR 30 million, investor holding under 5% β reveal that the drafters understood this. But the thresholds are arbitrary and narrow. Innovation doesn't stop at year five. A company with EUR 35 million in revenue can still be pre-profit. And the investors most critical to a startup's survival β founders and early backers β almost always hold more than 5%.
The result is a selection effect. Capital migrates toward assets with predictable, modest returns β bonds, dividend stocks, real estate β and away from the high-variance bets that drive technological progress. The tax doesn't just redistribute wealth. It reshapes what gets built. A system that taxes every paper gain annually makes patient, concentrated, illiquid investment β the kind that produces the next generation of companies β systematically more expensive than parking money in index funds. In a world where most economic value is created by intangible assets that take years to mature, that's not a technicality. It's a structural tax on innovation.
The Scoreboard
The Netherlands joins a long and instructive list. Twelve OECD countries levied wealth taxes in 1990. Today, four remain: Norway, Spain, Switzerland, and Colombia. The others β Austria, Denmark, Finland, France, Germany, Iceland, Luxembourg, Sweden β all abolished theirs, usually after discovering that aggressive taxation of wealth or unrealized gains drives capital across borders faster than it fills treasuries.
France's experience is the most thoroughly documented. The ImpΓ΄t de solidaritΓ© sur la fortune ran from 1982 to 2017. Over that period, an estimated 60,000 millionaires left France, taking approximately EUR 200 billion with them. The annual fiscal shortfall β revenue lost through emigration of taxpayers β was roughly twice what the tax collected. Macron abolished it in 2017.
Sweden's wealth tax survived from 1911 to 2007. Its most famous casualty was IKEA founder Ingvar Kamprad, who left in 1973 and didn't return for four decades. After abolition in 2007, wealthy Swedes began returning. The country subsequently produced Spotify, Klarna, and one of the world's highest billionaire-per-capita ratios.
Norway raised its wealth tax to 1.1% in 2022. By 2024, some 300 multimillionaires and billionaires had relocated, many to Switzerland. Overall wealth tax revenue still grew β the broad base of 655,000 taxpayers compensated for the departures β but the long-term effects on investment and entrepreneurship won't be visible for years.
The Constitutional Irony
The deepest irony of the Dutch case is that the government was forced into this by its own courts. On Christmas Eve 2021 β the Kerstarrest, or Christmas ruling β the Dutch Supreme Court declared the existing Box 3 system a violation of the European Convention on Human Rights. The old system taxed investors on a fictional rate of return, regardless of what they actually earned. During the zero-interest-rate era, savers holding bank deposits were taxed on assumed returns of 4-6% while actually earning 0.01%. Some paid more in tax than their total income from savings. The Court called this a disproportionate interference with the right to property.
The government's response was to replace fictional returns with actual returns β but then tax those actual returns even before they're realized. It traded one fiction for another. The old system pretended you earned money you didn't earn. The new system pretends you received money you haven't received.
The Court demanded taxation based on economic reality. What it got was a system that creates a different kind of unreality β one where a number on a screen that could vanish tomorrow triggers a tax bill today.
The Real Question
The Dutch experiment failed fast enough that the damage will likely be contained. The new coalition has committed to replacing the mark-to-market component with a conventional realized-gains tax by 2028. But the underlying tension won't go away.
Wealth is increasingly held in assets that appreciate without generating cash. The "buy, borrow, die" strategy β hold appreciating assets, borrow against them for spending money, let heirs inherit at a stepped-up basis β allows the very wealthy to live lavishly while reporting minimal taxable income. This is a real problem, and "just tax unrealized gains" is the most obvious solution.
It is also, as the Netherlands just demonstrated, the solution that doesn't survive contact with reality. The asymmetry problem, the liquidity problem, and the capital flight problem aren't technical glitches to be patched. They are structural features of what it means to tax something that hasn't happened yet. Fourteen OECD countries have now adopted exit taxes β taxing unrealized gains when a taxpayer leaves the jurisdiction β as a more targeted alternative. It's less satisfying than annual mark-to-market taxation, but it has the considerable advantage of actually working.
The Dutch parliament voted 93 to 57 for a tax on money that doesn't exist. Thirteen days later, the government admitted it couldn't work. Somewhere in that thirteen-day window is the entire history of wealth taxation in miniature: the political appeal of taxing paper wealth, followed by the economic reality that paper wealth is made of paper.
Links: Dutch Parliament Greenlights Box 3 Tax (NL Times) | New Cabinet Pulling Back Box 3 Plan (NL Times) | Box 3 Actual Return Act Adopted (Deloitte) | Netherlands New Box 3 Law (KPMG) | Supreme Court Box 3 Still Contrary to ECHR (Meijburg) | Dutch Government Bows to Pressure (IFC Review) | Norway Wealth Tax Exodus (Fortune) | Failure of Norway's Wealth Tax Hike (Brussels Report) | France Wealth Tax History (Fortune) | Moore v. United States (Tax Foundation)