The marginal hour, and why more work sometimes adds so little

This works through marginal earnings in the order the parts actually depend on each other.

Average and marginal are different questions

Your average rate describes what the whole week pays; your marginal rate describes what the next hour adds after everything responds to it. Decisions about overtime, an extra shift or a second job depend entirely on the marginal figure and not at all on the average.

The two can differ substantially, which is why extra hours sometimes produce far less than the rate implies. Working out the marginal figure once, for your own circumstances, is more useful than any general statement about it. Because it depends on your income, household and country, nobody can tell you the number without knowing all three.

Where support tapers

Means-tested support of any kind that reduces as income rises effectively claims part of every additional unit earned. Where several schemes taper at once over the same income range, the combined effect can absorb a large share of extra earnings. These interactions are rarely designed together, which is how unusually steep ranges appear in systems that look reasonable individually.

The specific thresholds, tapers and interactions are entirely jurisdiction-specific and are adjusted periodically. A government or advice charity calculator will model your own position far more accurately than any rule of thumb.

Costs that step rather than scale

Some costs do not rise smoothly with hours but jump at a threshold, such as an extra childcare session or a different travel ticket. An additional shift that triggers a full day of childcare can cost more than the shift earns. Parking, meals bought out and additional travel behave the same way, arriving in whole units rather than in proportion.

In practice, these stepped costs are frequently larger than the tax effects that get most of the attention. Listing them for a specific proposed pattern is a five-minute exercise that changes decisions.

Where thresholds sit in tax and contributions

Most systems have income points at which rates change, allowances taper, or contribution liabilities begin or end. Crossing one of these points affects only the income above it in most designs, which means gross pay still rises even when the marginal rate does. Some systems do contain genuine cliff edges where crossing a point removes something entirely, and those deserve real attention.

Put simply, the rates, points and mechanisms differ by country and change with each budget, so no figure here would remain accurate.

This is general information rather than tax advice, and a qualified adviser is the right source before restructuring your hours around it.

Non-financial costs of the extra hour

The marginal hour is usually taken from rest, family time or the things that make the other hours sustainable. Fatigue reduces output in the paid hours and raises the risk of the mistakes that cost money in other ways. Extra hours taken continuously also remove the capacity to look for better-paid work, which is the change that would actually help.

Those costs do not appear in any calculation and are frequently the largest ones. Weighing them explicitly is what distinguishes a decision from a drift into permanent overtime.

Running the calculation properly

Take a specific proposed pattern rather than a general question, since the answer depends on the exact hours. Model net income with and without it using a local calculator, and add the stepped costs the calculator will not know about. Divide the difference by the additional hours to get the honest marginal rate for that pattern.

In practice, compare that figure against what the same hours could produce elsewhere, including in a better-paid job. Where the marginal rate is poor, the answer is usually a different job rather than more hours in this one.

The takeaway

Model the exact pattern, add the stepped costs, and divide by the extra hours before agreeing to anything.

Small and repeatable beats ambitious and abandoned, almost every time.