The thinning out of the labor market middle

The US has gained 387,000 managers and lost almost 2m clerical jobs since 2007, as new technologies replace office workers and plunge the American middle class deeper into crisis.

Data from the Bureau of Labour Statistics divide the US workforce into 821 jobs from dishwasher to librarian. They show rapid structural shifts – on top of a cyclical unemployment rate of 7.7 per cent – that may increase income inequality.

One probable cause of rising inequality is new computing technologies that destroy some middle-class occupations even as they create jobs for highly skilled workers who can exploit them.

The number of clerical workers such as book-keepers, tellers, data entry keyers, file clerks and typists has been falling, pointing to a structural decline. The number of retail cashiers has also dropped – indicating that internet shopping and self-checkout systems may be eroding another occupation.

Employment growth came from healthcare, management, computing and food service jobs. The number of personal care aides is up 390,000 since 2007. Demand for people who figure out how to replace clerical workers – such as operations managers, management analysts and logisticians – grew substantially.

…But salaries for many of the fast-growing occupations are lower than those they are replacing. The average wage for a clerical job in 2012 was $34,410 compared with $24,550 for a post in personal care. The average computing wage was $80,180 and $108,570 for managers.

The FT story is here.

I would add this.  It is often a mistake to think that “cyclical” and “structural” explanations of unemployment are conceptually separate.  When there is a liquidity crunch, as there was in 2008-2009, business owners must plan rather rapidly for the future, and decide which capabilities they will hold on to and build up, and which they will let go of and let rot.  It is as if a big part of future plans is suddenly compressed into current decisions.   Some unemployment will result, but it is both cyclical and structural at the same time.  A separate point is that the “structural” features will help determine how much of a nominal wage decline may be required to maintain or restore employment.  Furthermore in many search models the distinction between structural and cyclical unemployment is also not easy to define.