In his State of the Union speech, President Trump said, “Since the election, we have created 2.4 million new jobs, including 200,000 new jobs in manufacturing alone.” Is the latter good news?

Politicians seem to have a fetish for manufacturing. But economists tend to be unconcerned about the composition of employment across sectors.

Proponents of the virtues of manufacturing believe governments should actively seek to encourage it, as manufacturing has historically enjoyed fast productivity growth. They bemoan the large fall in manufacturing employment since 1979, arguing that if more resources and employment had been directed to the sector, faster growth and higher living standards would have resulted.

A new paper by Robert Lawrence suggests this reasoning is precisely backward. His evidence suggests there is a direct tradeoff between manufacturing employment and productivity growth.

Suppose a new innovation trickles through the manufacturing sector, raising the productivity of workers. This increases supply. The extent to which these feeds through to higher output or lower prices depends on the elasticity of demand – i.e. the slope of the demand curve.

Lawrence suggests that for the manufacturing sector as a whole, demand is not very responsive to price changes. As prices fall in part due to fewer workers being needed to produce the same output, demand does not really change. Consumers pocket the savings and spend more on services. This is compounded by workers spending relatively more on services from higher incomes resulting from the productivity improvements.

As a result, between 1947 and 2017, the share of consumer spending going to goods has fallen from 62 to 33 percent. Similar declines in manufacturing employment have been seen across developed economies. And this is robust to trade patterns. Lawrence outlines “985,000 US manufacturing jobs estimated to have been lost due to Chinese imports between 1999 and 2011 represents less than a fifth of the total loss of over 5 million US manufacturing jobs over the same period.”

Why does this matter for Trump’s tweet? Well, because the uptick in manufacturing employment since 2010 actually has come during a period of slow manufacturing productivity growth. Between 2010 and 2016, output per full-time employee in manufacturing fell by 2.2 percent vs. growth of 4.3 percent between 2000 and 2010. Manufacturing prices actually rose relative to a general GDP deflator during this period. Given manufacturing demand is largely fixed and unresponsive, worse productivity means the need to employ more workers to keep a steady output.

This implies two things. First, that apparent manufacturing employment success may merely be a signal of a less innovative sector with weakened productivity. Second, even if government initiatives were successful in improving manufacturing productivity, this would not generate significant blue-collar employment.