Unemployment became a big problem in theoretical economics with Keynes. Economists had certainly thought about it before, but Keynes insisted that traditional approaches to unemployment were risibly inadequate and unemployment must be front and center in economic theory.
But is the concept of unemployment adequately defined, philosophically speaking? In particular, are economists correct in identifying the unemployment figures reported by the BLS with the economy being below the production possibilities frontier?
Above is an example taken from my slides for ECON 105 of a production possibilities frontier with a point representing unemployment. In the usual style of comically simple textbook examples, it assumes there are only two goods, industrial robots and pizzas. Of course, the model could be extended to n goods, only it couldn't be represented graphically in that case, only mathematically. Point U represents a situation of unemployment. That's one definition of unemployment. Call it unemployement in Sense A.
The other definition of unemployment is that calculated by the BLS: people who are actively looking for work but can't find a job. Call that unemployment in Sense B. It tends to be casually assumed that these two definitions are roughly synonymous.
But it's easy to imagine situations where the economy has unemployment in Sense A but not in Sense B. Suppose Bill is a top-notch computer programmer, and a pretty good entrepreneur, too, but he's working at McDonalds for $10/hour. He could be producing $100K; he is producing $20K. Bill is unemployed in Sense B, but not in Sense A. Call this the Diamonds-in-the-Rough case.
It's also possible that the economy has unemployment in Sense A but not in Sense B. Suppose Jack is looking for a job, but there is really nothing he can do that anyone would be willing to pay him for. He is a "zero marginal product" (ZMP) worker, to use a concept advocated by Tyler Cowen. Jack is unemployed in Sense B, but there is no unemployment in Sense A, since Jack is not an economic resource. What is confusing here is that technological change can render a person economically useless who formerly had value, and whose skills are intact. Jack might be great at doing sums, for example, but is rendered useless by the availability of various calculating devices (smartphones, computers, etc.). I am less interested in the ZMP case, however, than in the Diamonds-in-the-Rough case.
Unemployment in Sense B may be a social problem, in the sense that people's self-esteem depends on having a job. That wasn't always the case: there have been prestigious leisure classes in history. Whether the stigma we attach to unemployment is good or bad is an interesting question. It could be inefficient, if people accept lousy jobs to avoid the stigma of unemployment, when a bit more search could raise their marginal product substantially.
Unemployment in Sense A is a problem in a truer sense: it represents a real social loss, an inefficiency. I wonder if we worry too much about unemployment in Sense B, and if that concern is self-defeating, because the best way to keep down unemployment in Sense B is to keep unemployment in Sense A down. If the Diamonds-in-the-Rough realized their full value, they'd create a lot more jobs for others.
Sorry, this is all quite vague. Just brainstorming here.