CliffCheckYour data stays on this devicePrivate, on-device

Why this happens

How a raise can leave you poorer

At $47,000 in Ohio, a $1,000 raise carries an effective tax rate ofover 100%The extra pay is wiped out by lost benefits and higher costs, so the family ends up $19,076/yr poorer ($1,590/mo) than before the raise. It is not your mistake.

A rate over 100% means the raise leaves you with less than you started with. It happens because several benefits fall away in the same narrow band of income, all at once, faster than the extra pay comes in. The programs were each built on their own, decades apart, so no one designed the collision. This page explains the math, the history, and what can be done about it.

Check your own number

The math

Your real tax rate is not the number on your payslip

Economists call it the effective marginal rate: for each extra dollar you earn, how much you actually keep after tax and after any benefits you lose. On a payslip your marginal rate might be 20 or 30 percent. But benefits like food help, health coverage and childcare support are tied to income, so as your pay rises they shrink. Count both the tax you pay and the benefits you lose, and the real rate on a raise can pass 100 percent. When it does, earning more leaves you with less.

The worked example, a family of 4 in Ohio

The raise, before tax+$1,000
Benefits and extra costs it triggers−$19,076
What the family actually keeps−$18,076

Losing $19,076 to gain $1,000 is an effective tax rate far above 100%. The figures are calculated live from published government rules for this household, the same engine behind the calculator.

The Atlanta Federal Reserve built a research tool, called CLIFF, to map exactly these moments for career counselors. The core finding is consistent across studies: in certain income bands, working families face effective rates far above what any top earner pays.

The history

No one designed the cliff. It emerged.

Each of these programs was created on its own, to solve one problem, often decades apart. None was designed to work alongside the others, so each sets its own income limit, and where several limits land close together, the drops stack into a cliff. The Center on Budget and Policy Priorities documents how the rules for a single program alone run to dozens of pages, before you stack five of them.

SNAP, food helpCreated in the 1960s to fight hunger. Its income limit was never set to line up with health or housing programs.
Medicaid, health coverageBuilt in 1965 for a different population, later widened, so its cutoff sits at yet another income point.
ACA marketplace savingsAdded in 2010 to fill the gap above Medicaid, with its own separate formula for how savings taper as pay rises.
Section 8 housingA voucher program tied to local rents, with a limit that has nothing to do with the food or health cutoffs.
Childcare helpRun state by state, each with its own entry and exit income, and often the steepest single drop of all.

Stack five limits that were never meant to meet, and a narrow band of income appears where a family can lose several supports at once. That band is the cliff. It is an accident of history, not a decision anyone made.

The asymmetry

It hits hardest exactly where a raise should matter most

The cruel part is where the cliff sits. It lands on working families in the income range just above the poverty line, the households for whom a promotion or a few extra hours is supposed to be the way up. A high earner who gets a raise keeps most of it. A parent leaving one benefit band can keep none of it, and sometimes less than none.

Research from the Urban Institute and the Brookings Institution finds the highest effective marginal rates in the country fall not on the wealthy but on low and moderate income households moving up. The people most told to just take the promotion are the ones the math punishes for it.

What can be done

There are fixes, at the policy level and for you today

None of this is inevitable. It comes from limits that were never lined up, so lining them up is the fix. Several approaches already exist and are used in some states:

Broad based categorical eligibility, or BBCEA rule most supported states already use that raises the food-help income limit above the federal default, so the drop is gentler.
Gradual step-downsPhasing a benefit out slowly as pay rises, a few dollars lost per dollar earned, instead of everything ending at one income line.
Earned income disregardsRules that let a family keep some of a benefit for a while after a raise, so a new job does not trigger an instant loss.

These are neutral descriptions of tools already in use, not a recommendation. Which ones a state adopts is a policy choice.

For your own situation today, there is one lever you may control. Money you put into an HSA or a pre-tax 401(k) lowers the income figure benefits are measured against, which can keep you under a limit while still building savings. The calculator has an Advanced levers panel where you can try this against your own numbers.

Try the pre-tax levers in the calculator

Keep going

See how the numbers are sourcedThe Ohio cliff in detail