Dutch box 3 taxes: the January 1st number your tax return wants
Every spring, the same ritual: the Dutch tax return asks what your investments were worth on January 1st, and you go digging through broker statements from fifteen months ago.
It is a strange kind of homework. The number itself is simple, what was everything worth on one specific morning, but the reconstruction is not, because brokers report it in different places, in different formats, and sometimes not at all. This piece walks through why the date matters, where each type of broker puts the answer, and how to make next spring's version of you grateful to this year's version.
Why January 1st
Box 3 has traditionally taxed assets on a single reference date, the peildatum, of January 1st. Not your average balance, not your year-end after the January dip: the value at the start of the year. Assets are categorized by what you held on January 1st, cash you invest on January 2nd still counted on the 1st, just as savings in a lower bracket. And anti-arbitrage rules (peildatumarbitrage) block selling in December to rebuy in January.
The logic of a single reference date is administrative, not philosophical: it gives the tax authority one snapshot to assess instead of a year of movements. The mechanics work roughly like this: the assets you held on the reference date are placed into categories, each category is treated according to its own rules, and the resulting figure feeds into your assessment. The exact rates, brackets, and exemptions attached to those categories change from year to year, sometimes substantially, so this article deliberately avoids quoting any. For the figures that apply to the year you are filing over, the Belastingdienst website is the only source worth trusting; anything you read on a blog, including this one, ages faster than milk.
What does not change is the structure of the question: the return wants to know what you held, and what it was worth, on one particular date. That is a bookkeeping question, and bookkeeping questions reward people who kept the books.
Where brokers hide it
DeGiro emails a yearly report (usually late January) with the January 1st value; you can also reconstruct it from a portfolio snapshot. Interactive Brokers needs a statement dated over year-end. Trading 212 and most app-brokers: no dedicated report, you're reconstructing from history. Multiple brokers means doing this dance several times, then adding it up yourself.
Prefilled tax returns increasingly include Dutch broker values, but foreign brokers (IBKR, T212, eToro) frequently arrive blank or wrong, and checking a prefilled number still requires knowing the real one.
The pattern is worth noticing. Brokers with a Dutch heritage tend to understand what their customers need in April and produce a document for it. International platforms serve dozens of tax regimes and, reasonably enough, serve none of them particularly well; they give you raw history and wish you luck. Neither is doing anything wrong. It just means the burden of producing one clean number lands on the person with the least tooling for it: you.
The prefill trap
A word on prefilled returns, because they create a specific kind of false comfort. When the tax authority prefills a value from a Dutch bank or broker, most people accept it without checking, which is usually fine. The trouble starts with accounts that arrive blank, or, worse, arrive wrong: a foreign broker's data that was matched incorrectly, or a value that covers one account but not another at the same institution. You remain responsible for the accuracy of your return either way. Verifying a prefilled figure requires independently knowing the right one, which puts you straight back at the reconstruction problem the prefill was supposed to spare you.
The reconstruction problem
Your portfolio value on January 1st = every position you held on December 31st × its closing price × that day's FX rate. For a multi-broker, multi-currency portfolio, that's an evening in Excel that everyone puts off until the deadline week.
Consider what the evening actually involves. First, establishing your holdings on December 31st, which means replaying every buy and sell up to that date, per broker, and hoping you did not miss a stock split or a corporate action along the way. Second, finding a year-end closing price for each position, including the delisted fund you sold in February and no longer see in any app. Third, converting anything priced in dollars or pounds at the year-end exchange rate, not today's. Fourth, doing all of it again for the broker account you almost forgot, the one with the leftover free share. Each step is easy; the accumulation is why the task gets postponed until deadline week, and why deadline week is unpleasant.
If you hold assets across currencies, the FX layer deserves particular respect: a portfolio's euro value on January 1st depends on that day's rates, and eyeballing it with current rates quietly shifts the number. We cover the general problem in multi-currency portfolio tracking; at tax time it stops being abstract.
Doing it by construction instead
A tracker that stores your full history does this by construction. BullBenchmark's year report shows the January 1st value for every year you've invested, computed from your own transactions and historical prices, all brokers combined, in euros. Come tax season, the number is a glance, not a project.
The underlying idea is simple: if a system knows every transaction you ever made and can look up historical prices and exchange rates, then your portfolio's value on any past date is a computation, not an archaeology dig. January 1st is just one date among thousands the system can already answer for. The same property that lets a tracker draw your portfolio chart over five years produces your tax snapshot for free. People with multiple brokers feel this hardest, since the alternative is repeating the reconstruction once per platform and summing the results by hand.
Why full history beats the snapshot anyway
One more reason to keep your full history: since the tegenbewijsregeling (the actual-return election introduced after the Supreme Court rulings), you may want your real yearly return, start value, end value, deposits, not just the January 1st snapshot. That's begin- and end-of-year values per year, which is exactly what a full-history tracker computes anyway.
The broader principle: tax regimes change, and the direction of travel in the Netherlands has been toward caring about what your investments actually did over the year, not only what they were worth on one morning. Whatever the rules look like by the time you file, a complete transaction history is the raw material from which any of these figures can be derived. A single saved screenshot of your December balance is not. Keeping the history is the cheap insurance; the expensive alternative is reconstructing it under deadline pressure, annually, forever.
A practical year-end routine
For those who like their rituals explicit, the low-effort version looks like this. In the first week of January, download a fresh export from each broker, while the year-end is recent and the export screens are familiar. Upload them to your tracker, or at minimum file them somewhere findable. If a broker offers an annual tax report, save it when it arrives, DeGiro's, for instance, tends to show up in late January. Then stop thinking about it until spring, when the number you need is already sitting in your year report instead of in fifteen months of statements. Total time investment: about ten minutes, once a year, in exchange for never doing the Excel evening again.
Box 3 is mid-reform and the rules shift almost yearly, for actual advice, talk to an adviser. We do data.
The first week of BullBenchmark is free, no card, upload the export you just downloaded and see where you stand.
Related: DeGiro export guide