Presidential hopefuls across the political spectrum from Donald Trump to Bernie Sanders want to blame global trade and specifically “bad trade deals” for the ongoing loss of US jobs, particularly in the manufacturing sector. They say that better US negotiating strategies for such free trade agreements (FTA), or possibly eliminating them altogether, will ultimately restore the US economy to the high wage manufacturing powerhouse it was in the 1940s, ‘50s and ‘60s. This is highly unlikely at best and is a completely misleading distraction to boot. Yes, FTAs like the TPP and NAFTA were bad for American workers but there were other more important causes for the decline in average US living standards. In addition, manufacturing has been making a comeback to the US economy. As early as 2012, President Obama credited a precipitous decline in unemployment levels in domestic manufacturing as evidence of that sector’s contribution to the US economic recovery.
Manufacturing is making a return but unlike the US manufacturing sector in its heyday, today’s is very low wage, often paying below average levels in the overall economy. One 2014 study shows that roughly half the 6.2 million non-supervisory production workers in the US manufacturing sector earn well below $16/hour easily putting them at the federal poverty level for a family of four. Unlike in prior decades, one earner with a manufacturing job can’t support an average size family. Low real wages is one big reason the study shows a marked trend in “onshoring” of US manufacturing jobs since the start of the recovery. It's for this very reason that no one should expect the gradually recovering manufacturing sector to restore the US middle class to its previous living standard.
Let’s look more closely at the correlation between net job loss and international trade. Well known labor economist, Stephen Rose, has shown that the relationship is not as strong as the media pundits would have us believe. He points out that in the from 1940s, when US manufacturing employment peaked at 35% until the recession of the mid-1970s when employment in that sector dropped to one fifth, the US balance of trade deficit remained strongly positive. In other words, imports played little role in the dramatic loss of the manufacturing sector’s share of employment in the US economy. True, manufacturing’s share of employment continued its long term decline after NAFTA took effect in the mid-’90s, but not nearly as steeply as during the earlier period of high growth and low unemployment in the early post WWII decades. Slow overall economic growth driven by low levels of consumer demand and real wage growth were likely greater culprits than imports. In addition, higher productivity due to efficiency from technological change is a much greater contributor to manufacturing’s decline in employment share. As one paper in the Harvard Business Review recently confirmed, even though trade deficits as a share of GDP have been growing more rapidly since the late ‘90s, so has output per worker “...so the job content of each dollar of deficit has been falling rapidly” making goods imports alone less and less culpable for falling levels of domestic US manufacturing employment. The paper goes on to state that average annual rates of decline in manufacturing’s share of national employment from the end of the Clinton Administration until the start of the most recent recovery in 2010 are almost the same as during the first thirty years of the post WWII era and that this phenomenon has been experienced by all the advanced capitalist economies.
The argument that “labor market friction” (unemployment due to temporary adjustments such as job changes often due to the negative impact of foreign trade) is based heavily on trade’s differential effect on different sectors so that comparing long term trends in manufacturing employment and the overall trade balance tells us little about a possible correlation. If US exports (in agriculture and services for example) remained high as a gradual increase in manufacturing imports occurred than the a positive correlation could exist. Yet as Paul Krugman points out, despite a negative shift in the balance of trade in manufacturing goods against the US since the mid-1970s (there was a rough balance before that-and even a surplus in our favor at times-all during the decline in the US manufacturing sector's employment share), manufacturing imports are not automatically the main culprit in the decline in US manufacturing employment. Krugman points out that since so much of what we now import contains US domestically manufactured components the raw balance of trade figures don’t tell the whole story, “...because not every dollar of manufactured exports (or imports) corresponds to a dollar of manufacturing value-added.” Global era supply chains are complex things to understand, measure an evaluate. This is one reason that tariffs won’t have the positive impact on our manufacturing economy we want them to have and why this facile solution has failed miserably so often in recent times; they could be shutting out some of our own domestically produced components. Thus, tariffs are likely to make things even worse.
It also needs to be pointed out that over the period of the ‘70s and ‘80s US manufacturing exports increased even as imports did as well. One article coauthored by Krugman shows that from1970 to 1990, US manufactured exports rose from 12.6 to 31.0 percent of value added. Krugman and Lawrence stress that the import impact on domestic manufacturing employment is exaggerated by the fact that value added, the total sales prices of domestic manufactured goods minus the purchases of the manufacturing sector from other businesses to make those goods, isn’t taken into account. Krugman conducted a study based on US Commerce Department data that showed that each dollar of goods imports reduced US manufacturing’s contribution to GDP by only 60 cents so that even had there been a perfect balance of trade in goods over the two decades under consideration, the decline of manufacturing’s contribution to US GDP would have still been 86% as large as it actually was with the growing trade deficit. This is because the displacement of domestic spending on manufacturing by imports is largely from the service sector or even from manufactured inputs from abroad! The deindustrialization process would have marched on with very little change in the impact on employment levels (or wages). Much of the displacement of manufacturing by imports represents “leakage” or value added purchased from non-manufacturing sectors or from manufacturers abroad neither of which directly affects domestic manufacturing employment.
Some economists may see this as irrelevant or perhaps even, a description of the real problem; labor economist Josh Bivens of the Economic Policy Institute (EPI), which holds the position quite opposite that of Krugman, argues that manufacturing jobs on average have the highest “respending and employment multipliers” meaning that each manufacturing job in the economy can support nearly three jobs in other sectors on average-”Each 100 jobs in manufacturing supports 2.91 jobs elsewhere in the economy”-and thus job loss in manufacturing have the greatest negative spillover effect for the economy as a whole. Bivens explains;
All job loss is not alike, in terms of its effects on the wider economy. The employment multipliers calculated in this paper support the proposition that layoffs in the manufacturing sector tend to have much larger spillover effects in terms of indirect employment loss than layoffs in other sectors. There are essentially two reasons for this: manufacturing production tends to require many more intermediate goods and capital equipment than do many other sectors...and manufacturing jobs pay relatively high wages that lead to larger re-spending and government employment effects. While there is substantial variation in employment multipliers within manufacturing (from 175 in apparel to 464 in automobile production to 904 in computer equipment and office machinery), manufacturing industries across-the-board support more secondary employment than the retail trade or business and personal service sector.
BIvens explains in this EPI working paper dated 2003, that higher average levels of chronic unemployment and the stubbornly slow job market recovery after 2001 was caused by the nearly three million manufacturing jobs lost over the course of the late ‘90s and early 2,000s. This has been the general argument of the global trade competition school of thought on US manufacturing job loss; From the 1970s until the late 1990s, the US manufacturing employment levels were relatively stable-though manufacturing’s contribution to employment share in the economy declined somewhat with the overall growth of the workforce. By the late 1990s, all of this began to change. EPI trade economist,
Robert Scott sums up this narrative best in a more recent report;
Between 1970 and 2000, manufacturing employment was relatively stable, ranging from 16.8 to 19.6 million, and generally remaining between 17 and 18 million...However, this relationship broke down in the early 2000s, a period of rapidly growing trade deficits. At that time, manufacturing employment began a prolonged collapse, falling to a low of 11.5 million in February 2010, and recovering by December 2014 to 12.3 million, where it has remained. Overall, manufacturing lost 5 million jobs between January 2000 and December 2014.
Scott notes that during this time the rate of increase in manufacturing productivity slowed significantly so attributing such sudden and massive job loss to technology is highly implausible. Scott blames high and rising trade deficits as a share of GDP.
But another writer,
Doug Henwood of the Left Business Observer (LBO) has another perspective on the relationship between job loss and productivity. He believes that productivity growth was still high despite the slow down in the rate of increase after 2000 and that in any case, the fall in consumer demand for manufactured output due to falling real income growth was the
real cause of a sudden precipitous job loss in US manufacturing. Henwood’s analysis is worth quoting at length;
Respectable Keynesians argue that the problem is demand, which is a cyclical argument more or less, while the productivity story is more structural. But maybe both sides are right in a way. Productivity is way up but wages are flat. Workers have seen little of the productivity revolution.
The canonical tech-driven productivity acceleration took hold around 1995. From then until the cyclical peak year of 2007, productivity growth averaged 2.6% a year—4.3% in manufacturing. Compensation, which includes fringe benefits (see graphic caption), rose 1.7% a year—and 1.5% in manufacturing. Direct pay overall rose 0.9% a year—0.3% for factory toilers. During the recession, productivity growth slowed, employment collapsed, and wages rose modestly. In the “recovery years” of 2010 and 2011, productivity growth—this time not from a tech revolution but from sweating a shrunken workforce harder—has resumed growing at a 2.6% pace (2.1% is the very long-term average, by the way). But both compensation and direct pay have fallen by an average of 0.3% a year.
Henwood goes on to argue that since the early 1980s, the long term trend in the recovery of corporate profits, after a long decline since the late 1960s, had been strongly under way interrupted only briefly by the dot.com crash in 2000. This was aided by productivity increases, real declines in average wage levels and tax cuts for the rich. Elsewhere, Henwood produces a graph showing especially strong trends in labor productivity growth in the early years of the post 2009 recovery only to fall back down after 2012 after which point, as Scott himself acknowledges, a significant recovery in manufacturing employment then begins! Overall, the case that productivity growth has contributed more to manufacturing job loss than imports is particularly strong, especially since the recent growth in manufacturing employment accompanies a slowdown in productivity growth and real wages. Which returns us to one of Henwood’s original arguments regarding Keynes; low effective demand reduces output growth since growing real wages are needed to sustain it.
One reason that progressives should be extremely wary at best of trade related arguments about manufacturing job loss is that it removes the issue of the capitalist system completely from one of its major consequences, high average levels of structural unemployment. As the original, comprehensive argument goes about unemployment being one of capitalism’s core historic tendencies;
But if a surplus labouring population is a necessary product of accumulation or of the development of wealth on a capitalist basis, this surplus-population becomes, conversely, the lever of capitalistic accumulation, nay, a condition of existence of the capitalist mode of production. It forms a disposable industrial reserve army, that belongs to capital quite as absolutely as if the latter had bred it at its own cost.
This quote from the first volume of Das Kapital by Karl Marx argues that as capital becomes more and more productive the labor force must invariably decrease. No wonder that since the mid-1980s recovery from a global recession all the advanced capitalist economies, including the US, experienced heightened labor productivity along with lower levels of manufacturing employment. This is one of capitalism’s ways of restoring and preserving the all important rate of profit, replacing workers with machines. From this trend a growing and restive reserve army of unemployed labor grows to eventually threaten the very stability of capitalism itself.
And it is this very trend which progressives must resist! We must demand a vast and ongoing public investment effort to create full employment, preferably in “green jobs” such as infrastructure repair and expansion, health care, education, mass transit, renewable energy and energy conservation. I have always found it very disturbing that the best minds of the North American labor movement attack trade agreements rather than try to counteract the core tendencies of the capitalist system itself. We should be demanding public investment for full employment rather than demanding tariffs to protect currently non-existent jobs that will never return. Tariffs worked two hundred years ago when the world economy was very different and would only be a disaster today. They would create massive global industrial overcapacity and even worse unemployment, not stimulate new domestic output and jobs in hoped for newly “protected” industries. Part of the political problem for progressives is that we really need to scrap the old model of “business unionism” in favor of a more forward looking vision of a socialist society whose core goal is full employment! The reserved army of labor is created by capitalism itself not by international trade agreement whether good or bad.