Child Support Is a Shared Cost, Not a Slice of One Salary

The question the formula is answering

Almost every dispute about child support is really a dispute about what is being measured. Income shares measures the cost of the children and then divides it; a percentage-of-obligor model measures the payer's capacity and takes a slice of it; the German table measures the child's need and lets income only choose a row. Those are three different questions, and they produce different numbers for the same family, which is why an amount that looks generous under one is derisory under another. Before arguing about the figure, find out which question your system is asking, because the arguments that move the answer are different in each case.

The income-shares answer has one property worth understanding even where it does not apply: it is not linear in income. In the Michigan schedule the combined obligation for two children is 38.10 percent of combined income at the lowest table level and 22.69 percent at the highest, with a marginal rate that falls from 37.92 to 16.16 percent as income rises. A family on twice the income does not spend twice as much on the children, and the schedule was built from expenditure surveys that say so. That is why the naive check — a quarter of the payer's salary — overstates at high incomes and understates badly at low ones.

Overnights are money, and the exponent says how much

Michigan's parenting-time offset is not a straight-line credit. It weights each parent's base obligation by the other parent's overnights raised to the power 2.5, which makes the credit for the first few extra nights very small and the credit near equal time very large. On the worked example, going from 90 to 128 overnights — 38 extra nights, more than a month — reduces the monthly order from $1,280.63 to $1,012.41, about $7 a month per night. Going all the way to 182 takes it to $289.73. The curve exists because a parent with a handful of nights adds almost nothing to their own household costs, while a parent with half the year is running a second full home.

The practical lesson survives the arithmetic. In any system, a change in the residence arrangement is a change in the money, and the two negotiations are the same negotiation. Agreeing a schedule without pricing it is how people end up with an order that no longer matches the life they are actually living. If your system prices time in steps rather than in a curve, find out exactly where the step is; if it prices time in a curve, ask the calculator what one more night a fortnight is worth before you concede or demand it.

The add-ons are where the arguments actually happen

Base support covers the ordinary costs of a child: food, housing, clothing, everyday transport. It does not cover child care, health-insurance premiums or unusual medical expenses, and in the income-shares model each of those is added separately and allocated by income share, not split down the middle. That distinction is worth holding on to, because it is the source of most post-judgment disputes. In the worked example, A's share of the ordinary-medical allowance is 62.5 percent of $400 a year, or $20.83 a month — small, but the same 62.5 percent applies to every nursery invoice and every orthodontic bill that qualifies.

There is a second structural point hidden here. Child benefit, where it exists, is not a bonus on top: it is credited against the child's need before anybody pays anything. German law is the clearest example — § 1612b BGB says the child benefit attributable to the child reduces the child's cash need, by half where one parent is discharging their duty by caring for the child, and in full otherwise. With child benefit at 259 euros a month under § 66 EStG, the table amount of 670 and 784 euros in the worked example becomes a payable amount of 540,50 and 654,50. Anyone comparing two countries' numbers without checking whether family benefits have already been netted off is comparing two different quantities.

The floor, and what happens below it

Every system has a rule for parents who cannot pay. Michigan's is explicit: a parent earning at or below the Low Income Threshold of $1,255.00 a month owes 10 percent of income and nothing more, and between that threshold and the point where the ordinary formula would produce a smaller figure there is a transition equation that phases the obligation in at 50 to 70 percent of each extra dollar depending on the number of children. Income below the threshold is also left out of the combined family income entirely, so it does not inflate the schedule figure that the other parent then has to share.

Below the floor, the honest observation is that no formula can produce money that does not exist, and that a nominal order which cannot be paid does more harm than a realistic one. It accumulates arrears, it triggers enforcement, and in several jurisdictions it can suspend a driving licence that was the way the payer earned. If your file is near the floor, the useful work is documenting income and living costs precisely rather than arguing about the schedule, because the schedule is not where the decision will be made.

How to prepare a file so the number is defensible

Whatever the system, four documents do most of the work. A twelve-month record of each parent's income including variable pay, because a formula fed with one good month produces an order that survives twelve bad ones. A calendar of overnights as actually lived, not as decreed, because that is what the offset measures. An itemised list of the child's real costs — nursery, insurance, medical, school, activities — with invoices, because that is what turns an add-on from an argument into a line. And a note of every family benefit received, since the systems that credit it will do so whether or not you mention it.

Then run the calculation twice: once on today's facts and once on the facts you expect in two years. Orders are reviewable everywhere, but review costs time and money, and the version of the arrangement that survives is usually the one that anticipated the obvious change — a child starting school and leaving nursery, a parent returning to full-time work, an age bracket that steps up. Writing the review trigger into the agreement in advance is cheaper than litigating it afterwards, and it is the single most useful thing the calculator can help you draft.

Frequently asked questions

Does shared residence mean nobody pays child support?

No, and this is the single commonest misunderstanding. Equal time removes most of the transfer but not all of it whenever the incomes differ, because the child's standard of living should not depend on which house they are in this week. On the worked example, equal overnights still leave Parent A paying $289.73 a month, because A earns $5,000 and B earns $3,000. If the incomes were equal the transfer would be close to nothing. Time and money are two separate variables, and only one of them was equalised.

Which income counts — gross, net, or something else?

It differs by system and the difference is large. American formulas generally work from a defined "net" that starts with all income from all sources and subtracts a specified list of deductions, which is neither gross pay nor take-home pay. The French reference table uses the income actually received. The German table uses an adjusted net income from which work-related expenses and certain debts have already been deducted before you look up the row. Whichever applies, the operative word is "defined": the figure comes from a list in the manual, not from a payslip, and putting the wrong one into a calculator produces an answer that is wrong by more than any argument you were planning to have.

Do I still pay after the child turns eighteen?

Very often yes, and majority is rarely the switch people expect. Support in most systems follows education and self-sufficiency rather than a birthday. The German table has a fourth age column for children of eighteen and over and sets a separate need for a student living away from the parental home, currently 990 euros a month including up to 440 euros of rent — and it says explicitly that health-insurance contributions and tuition fees are not included in that figure. Portuguese law is equally direct: article 1905(2) of the Civil Code continues the obligation past majority until the young person turns 25, unless training is completed sooner. Check the rule in your own system, because the date it turns on is almost never the eighteenth birthday.

Can we just agree a figure between ourselves?

You can agree it, and in every system here an agreement is the normal and preferred route — but an agreement that is not approved, homologated or recorded in the form your law requires is usually not enforceable, and unenforceable is exactly the property you do not want in a promise that has to last fifteen years. Two more cautions. First, most systems will not let parents contract away a child's minimum entitlement, so an agreement far below the floor may simply not hold. Second, put the review mechanism in writing at the same time: what happens when incomes change, when a child changes school, when the residence arrangement moves. The cheapest litigation is the one the agreement already answered.

The calculator uses one American state. Is that any use to me?

It is useful as a model of the mechanism rather than as an answer for your case. Michigan was chosen because its schedule is published as equations rather than a six-hundred-row lookup table, so the income shares, the parenting-time offset and the add-on allocation can be implemented exactly instead of approximated — which makes the sensitivities visible. Run it to see how much an extra overnight is worth, how the combined obligation flattens as income rises, and how much of the total the add-ons represent. Then take those questions, not those numbers, to whatever schedule actually governs your file.