Turnover gets treated as one number. It is actually two things added together, and they have completely different price tags.
The first is the turnover you can fix. Someone leaves over scheduling, or pay, or a manager, and those are problems with an owner and a budget attached to them. The second is the turnover built into the job itself, and no operator fixes that in a quarter, or at all.
The number on the page does not tell you which is which. That is the whole problem, and it is why buyers discount for turnover and owners apologize for it while almost nobody asks the question that separates them.
In this episode we work through the arithmetic on a home care aide's pay. The hourly rate says one thing, the annual earnings say another, and the roughly ten thousand dollars between them is not a wage story at all. It is an hours story — and once you see that, you know which kind of turnover you are looking at.
Full transcript, with every figure and the document it came from:
Figures cited in this episode
Median hourly wage of seventeen dollars and thirty-six cents in 2024, median annual earnings just under twenty-six thousand dollars, forty-six percent working part time under thirty-five hours and forty-eight percent full time year round — PHI, Direct Care Workers in the United States: Key Facts 2025.
4,677,100 home health and personal care aide jobs, eighteen percent projected growth, median pay of seventeen dollars and twenty-one cents an hour — U.S. Bureau of Labor Statistics, Occupational Outlook Handbook.
Caregiver turnover of approximately seventy-five percent, the lowest in five years — Activated Insights, 2025 Benchmarking Report.
An aggregate Medicare payment decrease of 1.3 percent for CY 2026 — CMS Fact Sheet, CY 2026 Home Health Prospective Payment System Final Rule (CMS-1828-F).
A recommended seven percent base rate reduction for 2027 — MedPAC, March 2026 Report to the Congress, Chapter 8, Home Health Care Services.
About Ada
Ada is an AI co-host. Not a person, and not pretending to be one. She brings the data, she argues, and she is right at least once an episode.