Tax numbers are where a finance site can hurt people, so this page documents exactly how every worked tax example on Financial Expert Class is computed: where each rate comes from, what the examples deliberately simplify, when they get re-checked, and what happened the last time a rate changed underneath us. The consolidated country table lives on the EU ETF tax dataset page; this page is the method behind it.
Who does this
Me, Kestutis Balciunas. I am not a tax adviser in any jurisdiction. I am a Lithuanian tax resident who has filed his own investment taxes for over a decade, and I write about other countries’ regimes as a careful outsider working from primary sources. That distinction is stated here once so every article does not have to earn it separately: the examples are educational arithmetic, not advice, and your accountant outranks my worked example every single time.
The five rules every tax example follows
- The rate comes from the national tax authority or the law itself — never from another blog, and never from an AI model’s memory. If I cannot trace a rate to a primary source, the example does not get written.
- Every example is labelled with its tax year (“2026 rules”), because Budgets change rates — see the Ireland case below.
- Simplifications are stated, not hidden. Unless the article says otherwise, examples assume a single resident taxpayer, no church tax or local surtaxes, no marginal-threshold effects, and EUR figures. Where a threshold matters (Germany’s Sparer-Pauschbetrag, Lithuania’s €500 exemption), it is applied explicitly and visibly.
- The arithmetic is shown step by step so a reader — or their adviser — can re-run it and catch my mistakes.
- Fund-level withholding is kept separate from investor-level tax. The ~15% US treaty withholding inside an Irish-domiciled UCITS is structural leakage at fund level; it is never mixed into the personal tax calculation, because you cannot credit it.
A worked example of the method (Germany, 2026)
This is the standard the articles hold themselves to — the German Vorabpauschale on a €50,000 accumulating equity UCITS position at the start of 2026, using the published Basiszins of 3.20% (BMF, 13 January 2026):
| Step | Calculation | Result |
|---|---|---|
| 1. Basisertrag rate | 70% × 3.20% Basiszins | 2.24% |
| 2. Basisertrag | €50,000 × 2.24% | €1,120.00 |
| 3. Cap check | Cannot exceed the fund’s actual gain that year | assume not binding |
| 4. Teilfreistellung (equity fund) | €1,120.00 × 70% (30% exempt) | €784.00 |
| 5. Tax at 26.375% | €784.00 × 26.375% | ≈ €206.78 |
Every number in that chain is traceable: the Basiszins to the Bundesbank, the 70%/30%/26.375% parameters to German law, and the mechanics to the full Vorabpauschale guide. That traceability — not the answer — is the methodology.
Data sources by country
Rates and rules are checked against the national tax authority for each country covered, for example: BZSt and the Federal Ministry of Finance (Germany), Revenue (Ireland), Belastingdienst (Netherlands), Agenzia delle Entrate (Italy), podatki.gov.pl (Poland), Agencia Tributaria (Spain), Skatteverket (Sweden), Skattestyrelsen (Denmark), Vero (Finland) and VMI (Lithuania). EU-level rules (UCITS, PRIIPs, MiFID II) are checked against ESMA and EUR-Lex. Where a country’s authority publishes poorly in English, the article says the interpretation is mine and flags the uncertainty instead of faking confidence.
Data cut-off policy: what happens when a Budget moves a rate
Each tax article and each row of the country dataset states the tax year it describes. When a government changes a rate, the change is applied to the affected articles and logged publicly. The working example: Ireland’s Budget 2026 cut the ETF exit-tax rate from 41% to 38% effective January 2026, and the site’s Ireland guides were updated and verified against Revenue’s figures — that entry sits on the corrections log. A visible Reviewed date on a tax article means a logged human re-check of its figures happened on that date, per the editorial process; no log entry, no badge.
What the examples deliberately are not
- Not personalised. Marginal rates, allowances already used, joint assessment, regional surtaxes and residency edge cases change real outcomes. The examples show the mechanism, not your bill.
- Not legal interpretation. Where a rule is genuinely ambiguous (the Netherlands’ Box 3 reform is the standing example), the article says so and describes the range of outcomes instead of picking one.
- Not tax planning. Wrappers like Poland’s IKE/IKZE or Hungary’s TBSZ are explained mechanically; whether you should use one is between you and a licensed adviser in your country.
Conflict of interest
Tax content carries no monetisation pressure in either direction: no government, accountant network or tax-software vendor pays this site. Revenue comes from broker referrals and the paid Substack, disclosed in full in the affiliate disclosure — neither depends on any country looking better or worse in a table.
Found an error in a tax example?
That is exactly the kind of report I most want. Email [email protected] with the article URL, the figure you believe is wrong, and the primary source that shows the current rule. Confirmed fixes are applied to the article and recorded on the public corrections log — the Ireland entry there shows what that looks like in practice.