It used to be that if you’re the type of person who always has a job, you’re the type of person who can always find a job.
But a new analysis of job-finding rates complicates that picture. The study from the Federal Reserve Bank of Richmond examines three types of workers and how they’ve fared in their job search.
“Primary workers,” who make up about 55% of the US labor pool, are “almost always employed,” per the analysis. They have steady work histories and usually only show up in unemployment statistics when something has gone wrong.
Historically, primary workers had the best odds of finding a new job. But their job-finding rates have fallen 13 percentage points from a November 2022 high to a September 2025 low — the biggest drop for any group the analysis covered.
Secondary workers, who make up about 14% of the labor pool, are often out of work or cycling through different jobs. They often make up the bulk of measured unemployment, and have a historically lower job-finding rate. While those workers still have a harder time finding jobs than their more attached peers, they haven’t seen their fortunes fall as much; in the same period of November 2022 to September 2025, the job-finding rate for secondary workers fell just two percentage points.
This is new. In major recessions in the past, primary workers have seen their job-finding rates fall the most, followed by secondary workers. But now, primary workers are seeing that massive drop, while less secure secondary workers are not. So why would it be happening when the US economy is growing?
At the same time, the analysis finds that roles more exposed to AI — which the researchers define as the overlap of jobs’ tasks with the capabilities listed in AI patents — have the biggest drop in job-finding rates. Before 2023, when ChatGPT became publicly accessible, job-finding rates for roles across levels of AI exposure moved together. Since then, outcomes have worsened dramatically for the AI-exposed fields.
AI has also shaken the hiring landscape for the new workers who might otherwise become strongly attached to the job market: As the report notes, research has found that generative AI can “substantially increase productivity at the task level (particularly for less-experienced workers), potentially altering the composition of new hiring.”
For entry-level workers, that means a tighter labor market defined by roles requiring a wider variety of skills. For the experienced white-collar workers contending with a declining job-finding rate, that’s meant taking drastic measures — including pay cuts.
It’s not all bad news for white-collar workers. In the latest jobs report, both the professional and business services and information sectors added payrolls.
But those mild white-collar gains still weren’t enough to offset a troubled labor market. On the whole, the US lost jobs in July — another sign that job hunting might not get much easier anytime soon.
