How we work out our numbers

Every figure on this site should be checkable. This page explains where the data comes from, what we assume, and where we are estimating rather than measuring.

Last updated August 30, 2026

The principles

  1. Every number has a source. If we cannot point at where a figure came from, we do not publish it.
  2. Ranges beat false precision. Where the honest answer is "somewhere between", we say that rather than inventing a decimal place.
  3. We state the data year. Not the year we read something, but the year the underlying data describes. These differ more often than people expect.
  4. Conventions are labelled as conventions. The dollar-per-year-of-age allowance rule is a widely repeated custom, not a research finding, and we say so on the page.
  5. We say when a source is discontinued. The USDA child cost series has not been updated since its 2015 data year. Anyone quoting a current figure is adjusting old data.

Calculations run in your browser

Every calculator is client-side JavaScript. Nothing you enter is sent to a server, which is both a privacy property and a practical one: the calculators keep working regardless of what happens to our hosting.

How each calculator is built

Cost of raising a child

Starts from a birth-to-18 base derived from the USDA series, scaled by household income through piecewise interpolation between eight anchor points, then by a state cost-of-living multiplier. An only child carries a 1.27 multiplier because per-child spending falls as family size rises. Category shares come from the USDA proportions. College is excluded unless you opt in.

The country versions are genuinely different models rather than currency conversions. Germany applies a Bundesland factor reflecting how far each state has made Kita free. Japan scales only the education category by school track, at a documented ratio of about 3.3 times for an all-private path. The Global version abandons the base-total approach entirely and estimates a child's share of household consumption, which is the method used where no national study exists.

Daycare versus stay-at-home

Take-home pay minus annual childcare minus annual work costs. Break-even is total costs divided by one minus the tax rate. It is deliberately a short-term cash comparison, and the page says plainly that it omits retirement contributions, career progression, and re-entry difficulty, which usually push the other way.

Allowance by age

Age multiplied by a weekly rate, optionally rounded, with a 50/40/10 spend, save, give split. Both the rate convention and the split are customs rather than findings, and are labelled as such. There is no authoritative dataset establishing a correct allowance by age.

Screen time by age

Uses American Academy of Pediatrics guidance bands: avoid screens other than video chat under 18 to 24 months, and about an hour a day of high-quality programming from 2 to 5. Above age 6 there is no official hourly limit, so the figures shown are practical planning numbers and the page says so rather than implying an official standard exists. Typical usage figures describe reported behaviour, not a target.

Grocery budget

Anchored to the four USDA food plan levels, with an age factor for children and a household size adjustment reflecting that buying for more people is more efficient per head.

Stay-at-home parent salary

A replacement-cost method: weekly hours in each of nine roles, priced at median hourly wages, scaled by number and age of children and by region. We state the main weakness openly, which is that roles overlap in real life and the method does not net that out, so the blended hourly rate is more defensible than the annual total.

Diapers, feeding, and road trips

Straightforward arithmetic over quantities you supply, using a 30.4-day month. The assumptions, such as the flat wipes allowance and the lodging and food rates, are listed on each page.

What we do not do

Our sources

The full list, with publishers and data years, is on the sources page. The main ones are:

Corrections

If you think a number here is wrong, please tell us. We would rather fix it than defend it. Use the contact page, and if you can point at a source that would help.