[temp-dude-87844]

Unemployment numbers, as with most official government numbers, are based on a combination of surveys and self-performed actions, such as filing for unemployment, the latter of which essentially amounts to self-reporting [1]. These results are then aggregated into six metrics, U-1 through U-6 [2], of which U-3 has been designated the official and widely-quoted unemployment rate.

The main value of U-3 isn't that it's a particularly insightful reflection on job health, but rather that it's easy to measure, its definition is consistent, and captures a particular subset of the population that has demonstrated overt desire to participate in government programs that ostensibly help them on the path towards employment. This self-selects for workers exceeding a particular desperation ratio, while not counting people who may think that they'd prefer to work, but haven't taken a government-defined sequence of steps towards doing so.

Seasonally-adjusted U-3 is low right now, around 4.1%, so in macro terms the pool of people who have taken overt steps towards getting a job while still not having a job is shrinking. This means there's fewer candidates for a particular position, and on average, less of a chance that any particular candidate is qualified for any particular position. You could argue that this means there's a labor shortage.

[1] https://www.bls.gov/cps/cps_htgm.htm [2] https://www.bls.gov/news.release/empsit.t15.htm