Everybody seems to be worried about manufacturing these days. All the Democratic presidential candidates condemn the practice of “outsourcing” — laying off manufacturing workers and buying their output more cheaply from China. This is not surprising, given that organized labor has made it a high priority issue. But they are being joined by some on the right-wing fringe as well, such as Pat Buchanan and Paul Craig Roberts, who warn that we are exporting our sovereignty along with our jobs. They all seem to think that more trade protection is the answer.
The truth is that manufacturing is doing just fine in every way except employment. However, few economists would judge the health or sickness of any industry solely based on employment. By that standard, agriculture has been the sickest industry of all for decades. Rather, such things as output, productivity, profitability and wages better determine industrial health. On this score, manufacturing is actually doing quite well in the United States.
Let’s start with the bad news. According to the Bureau of Labor Statistics, there were 14.6 million Americans employed in manufacturing in July, down from 15.3 million a year earlier, 16.4 million the year before that (2001) and 17.3 million the year before that (2000) — a decline of 16 percent in 3 years. The recent peak for manufacturing employment occurred in March 1998 at 17.6 million — about the same as it had been for the previous 15 years.
By contrast, industrial production has remained relatively strong. The Federal Reserve Board’s industrial production index is up 5 percent since manufacturing employment peaked in 1998, and down just 5 percent from the index’s peak in July 2000, despite a rather severe recession in the meantime.
Looking at gross domestic product, real goods production as a share of real (inflation-adjusted) GDP is close to its all-time high. In the first quarter of 2003 — the latest data available — real goods production was 39.2 percent of real GDP. The highest annual figure ever recorded was 40 percent in 2000. By contrast, in the “good old days” of the 1940s, 1950s and 1960s, the United States actually produced far fewer goods as a share of total output. The highest figure recorded in the 1940s was 35.5 percent in 1943; the highest in the 1950s was 34.9 percent in 1953; and the highest in the 1960s was 33.6 percent in 1966.
In short, manufacturing output is very healthy. There is absolutely no evidence whatsoever that we are becoming a nation of “hamburger-flippers.” We are producing more “things” than we have in almost every year of our history for which we have data. The decline in employment is, in effect, a good thing, because it means that manufacturing productivity is very high. That is also a good thing, because it means that employers can afford to pay high wages to manufacturing workers while still competing with low-wage workers in places like Mexico and China.
Remember, what really matters for employers is not absolute wages, but unit labor costs — how much the labor costs to manufacture a given product. If a U.S. worker is five times as productive as a Mexican worker making one-fifth as much, they are exactly equal from the point of view of a producer.
The best measure of comparative productivity levels is real GDP per employed person. According to the Bureau of Labor Statistics, in 2002 the United States continued to lead the world in this category. All U.S. workers produced $71,600 in output each (in 1999 dollars). The next highest country was Belgium, where each worker produced $64,100. Japanese workers — renowned for their productivity — produced just $51,600. Korean workers produced even less: $34,600 each. (There’s no data for China or Mexico, but both are probably far below Korea in terms of productivity.)
It is also important to note that virtually every other major country has seen declines in manufacturing employment. Between 1992 and 2002, U.S. manufacturing employment fell by 3.7 percentage points. In Britain, it fell 4.7 percent, in Japan it fell 5.2 percent, and in Germany it fell 6.1 percent. Only Canada and Italy showed any increase over this period.
Finally, it is important to note that much of what is going on here is not “real” in some sense, but definitional changes in job classifications. It used to be that big companies tended to do everything in-house, so people like janitors and accounts were classified as “manufacturing” workers simply because they worked for manufacturing companies. Over the years, such companies discovered that it was more economical to contract out such work. That is why “business services” is one of the fastest rising categories of employment in the United States.
Stanford economist Robert Hall recently told the Senate Finance Committee, “There is no sign in the data on output of the onset of chronic ill health in manufacturing.” All the hand wringing is simply unjustified by careful analysis.
Chart data: Goods Production as a Share of GDP
Decade —- Percent
1940s —- 34.6
1950s —- 33.9
1960s —- 32.7
1970s —- 33.4
1980s —- 34.9
1990s —- 37.2
2000s —- 39.4
