How we calculate, and how we test it

Assumptions we add where the sources are silent

Each is shown here and, where it affects your result, in the result itself.

1. A part sale reduces cost for the $50,000 test by the average cost of the shares sold. IRD has not said which method to use (research gap 3). Flagged in every result with a sale.
2. With IRD rates, the 1 April value uses the end-of-month rate for 31 March of the year before, the 31 March value uses that day’s end-of-month rate, and trades and dividends use that month’s mid-month rate (IR461 p13 accepts it as an actual rate).
3. The rolling 12-month average for an income year ending 31 March is IRD’s figure dated 15 March of that year.
4. Under the rolling average, the 1 April value uses last year’s rate, for CV (s EX 51(5)) and, on our reading, for FDR too.
5. The quick sale adjustment is shared between holdings in proportion to their peak holding amounts, only to work out each holding’s foreign tax credit.
6. NZ tax on one holding’s income is that income × your average tax rate (total income tax ÷ taxable income). IS 21/09 was not read; check with your accountant.
7. Box 17A shows the foreign tax you can claim as a credit after the limits, not the raw tax withheld.
8. Credits on FIF holdings together never exceed the NZ tax on the net FIF income, and none are allowed when comparative value reduces FIF income to nil; credits on dividends never exceed the NZ tax on those dividends. Capped credits are scaled back in proportion, so the IR1261 lines add up to box 17A.
9. US dividend withholding is 15% with a W-8BEN: the NZ and US tax treaty as amended by the 2008 protocol, Article 10(2), caps US tax on dividends to an NZ resident at 15% (5% for a company with 10% or more of the voting power) (US Treasury copy). Without the form, 30% may be withheld (IRS Instructions for Form W-8BEN). Only the 15% can be credited in New Zealand.
10. Your share of joint holdings scales every NZ$ amount for those holdings, including cost.
11. IR1261 lines are one per income type and country; if comparative value is reduced to nil, FIF income goes on one “Overseas” line.
12. The transitional resident window (to the last day of the 48th month after the month you became resident) and the 10-year test come from IR1024 (April 2026) and IRD’s transitional residents page (updated 17 June 2026). Applying for Working for Families ends the exemption (IR461 April 2026 p22 note 3).
13. RAM eligibility follows IR461 p17; whether a person who was a transitional resident can use RAM once that ends needs confirmation.
14. An optional apportionment of RAM gains by days resident is our assumption and needs confirmation.
15. The PIE comparison assumes the fund uses FDR on the same shares. Some funds use other methods.
16. In the PIE comparison, tax is paid out of the investment each year on both sides and dividends are reinvested.
17. In the PIE comparison, the $50,000 test is re-run every year on cost (amount invested plus reinvested dividends, less the average cost of shares sold to pay tax). Dividends are reinvested within the year, so a year’s highest cost is its starting cost plus its dividends; tax is paid by selling shares after the year.
18. 2026/27 tax brackets are taken as equal to 2025/26 (not confirmed on an IRD page; the annual rates confirmation sits in the lapsed bill).
19. Figures are rounded half up to the cent for display (Hatch’s sample report does the same).
20. Where FDR and CV are equal we show FDR; the tax is the same.
21. Tax is worked out on whole dollars: like the IR3, we drop the cents from taxable income (for example FIF income of $6,189.26 is taxed on $6,189). The FIF income and box 17B figures keep their cents.
22. We choose between FDR and CV by which gives the lower FIF income, as IR461 p14 puts it (CV may be used when it gives lower income), not by which gives less tax after foreign tax credits. In rare cases with large credits the other method could leave less to pay; ask your accountant.
23. With your own rates, one rate per currency is used for all trades and dividends in the year: an approximation of the actual rate on each day.