The first paper is a report by the IMF, "The U.S. Manufacturing Recovery: Uptick or Renaissance?" which reports on the unusual speed of recovery in manufacturing since the end of the great recesssion. This is attributed to the weakness of the dollar (which encourages exports), restraint on wage growth, and lower US energy costs (presumably due to low prices on US natural gas produced by the fracking boom).
The second is a paper in the Journal of Economic Perspectives, "US Manufacturing: Understanding Its Past and Its Potential Future" by Martin Neil Baily and Barry P. Bosworth. Baily and Bosworth note that manufacturing employment as a percentage of the US workforce has been declining steadily for 40 years. The authors produce the following graphic to illustrate this:
Note that the decline in employment (as a share of total employment) has been very consistent since the mid 1960s. What is interesting here is that, if you look at the absolute number of people employed in manufacturing, the numbers actually increase into the 1980s and then decline mainly in two steps, one small step down in the early 1980s and one precipitous drop in the last decade (See graph below). These downturns in absolute numbers give the impression of precipitous events rocking the industry, especially since 2001. However, the data presented by Baily and Bosworth above shows a more constant downward trend.
Why are the graphs so different? By looking at manufacturing's share of the work force, Baily and Bosworth are implicitly taking into account changes in the total workforce that affect the absolute number of people working in manufacturing. For instance, the FRED graph shows that the number of people working in US manufacturing grew about 20% in the 1960. Yet this was also when the baby-boomers started entering the workforce, Indeed, the first baby-boomers turned 18 in 1964 and Baily and Bosworths's red employment line begins its downward slide soon afterwards. Note also that, in the FRED graph, manufacturing employment has its ups and downs (which are mostly due to recessions) in the 1970s and 1980s but stays close to 18 million. However, this is a period in which the remaining baby-boomers enter the workforce and the percentage of the population entering the workforce climbs from about 58% to 63% largely due to women entering the workforce in greater numbers. Therefore, by expressing employment as the share of total employment, Baily and Bosworth's data weeds out the effect of recessions and changing demographics.
This is particularly important since, the apparent free fall in manufacturing employment from 2001 to 2010 shown in the FRED graph coincides with a precipitous drop in the percentage of the population in the total civilian workforce. In this period, the combined impact of retiring baby-boomers and the Great Recession pushed this percentage from a historic high above 64% to below 59% (essentially back to late 1970s levels). Baily and Bosworth's data implicitly takes this into account and shows that, despite the appearance in the FRED graph, the alarming 5 million employee drop in manufacturing employment is nevertheless in line with the general trend since the mid 1960s.
Another thing to note in Baily and Bosworth's graph is that, since 1960, manufacturing's share of GDP has remained quite steady. Thus, the above graph makes a point that I have been trying to make (perhaps not as well as they do) that manufacturing in the US has not declined, just the employment in it has.
Baily and Bosworth also find that this decline in share of employment in US manufacturing is equivalent to the average decline seen in the G-7 countries. So, whatever is driving the trend in declining employment in manufacturing, it is affecting the other most industrialized nations in equal measure.
The big point here is that, if you look at the number of people employed in manufacturing (the FRED graph), you see an accelerating downward trend starting in the 1980s and you look for causes in the 80s, 90s, and, especially, the 00s. This leads one to consider things like NAFTA and FTAs, globalization, and China's currency manipulation. However, theses factors seem less plausible as explanations for declining employment in US manufacturing when you see the decline as part of a longer term trend, as in the Baily and Bosworth graph.
No comments:
Post a Comment