The Caregiver Wage Gap: What Families Pay and What Aides Earn

How to read this

The family-side hourly rate is the published 2025 modeled monthly median converted at the survey's own basis of 44 hours a week, under cost model cost-model-v1.4.0. It is the same conversion used on the cost pages of this site, so the figure here matches the figure a family sees when they look up their own state.

Two national readings are published here on purpose. $18.48 and 2.09 come straight from the two medians printed above them, so the arithmetic on this page checks. $18.77 and 2.12 are the medians of the 50 state gaps and the 50 state multiples in the table below, which is the better reading of what a typical state looks like. Neither is the derived form of the other, and the state table carries the row-level figures either can be rebuilt from.

The two columns are not independent measurements. Our cost model uses this same wage series as one of its inputs when it spreads survey medians across geographies. The overall size of the gap is still meaningful, because the survey median and the wage median are separately published, but small differences between states should not be read as findings.

The multiple varies more than the dollar gap does, because both sides move together with local wage levels. A state where families pay a little more usually also pays aides a little more.

None of this describes a caregiver hired directly by a family, or one paid through a consumer-directed Medicaid program. Those arrangements sit outside both the agency rate and the surveyed wage.

Methodology

The family side is our published modeled monthly median for in-home care, converted to an hourly rate at the survey's own basis of forty-four hours a week. That is the same conversion the cost pages on this site use, so the number here matches the number a reader sees elsewhere.

The worker side is the Bureau of Labor Statistics Occupational Employment and Wage Statistics annual median for home health and personal care aides, metro-weighted to the state, converted to hourly on a standard two thousand and eighty hour year.

Inside each state the gap is the difference between the two rates and the multiple is the family rate divided by the aide wage. Every state row publishes both alongside the two rates it came from, so a reader can check the arithmetic row by row.

The national figures are taken across states, and there are two honest ways to take them, so both are published. One subtracts and divides the two national medians printed beside each other, which is the reading a reader checking the page will reach for. The other takes each state's own gap and multiple first and then reports the median of those, which is the better description of a typical state. They do not agree, because a median of differences is not the difference of medians, and the page says which is which rather than presenting one as the other.

Sitewide cost and data method: Methodology · Data sources · Corrections

What this cannot tell you

The two sides are not independent. Our cost model uses the same BLS wage series as one of its inputs to spread survey medians across geographies, so part of the correlation between the columns is built in by construction. The size of the gap is still meaningful; the fine detail of how it varies between states is less so.

The difference between the two figures is not agency profit. An agency rate has to carry payroll taxes, workers compensation, liability insurance, scheduling and supervision, recruiting, training, background checks, unbilled travel, and cover for a caregiver who calls in sick. What is left over is a much smaller number than the gap.

The wage series covers home health and personal care aides employed across all settings, including agency employees, and is metro-weighted. Rural wages are represented only through the metro areas the series publishes.

The family-side hourly rate is derived from a monthly survey median at a fixed weekly-hours basis. A family buying twenty hours a week will usually face a higher hourly rate than one buying full-time care, so a single hourly figure flattens a real difference.

A caregiver hired directly, or paid through a consumer-directed Medicaid program, is not paid the agency rate and is usually not paid the survey wage either. Neither column describes that arrangement.

The national figures are summaries across states, not a national measurement. Neither reading is weighted by population, by how many families buy care in a state, or by how many aides work in one, so a small state counts as much as a large one. A weighted national figure would land somewhere else, and we do not publish one because the survey behind the family side is not published at the resolution that would need.