In a recent interview, Steve Bannon — Donald Trump’s onetime campaign strategist — outlined his political thinking in the run-up to the 2016 presidential election.

After Mitt Romney lost to Barack Obama in 2012, Bannon noted in the interview that international trade was “no. 100” on the list of voter priorities, while immigration ranked near the top of the list. Bannon opined that if he could find a way to weave together a storyline that connected international trade, immigration and working-class wage suppression, he would have the basis for an explosive campaign.

All he would need then was a powerful orator to deliver the message. He had Donald Trump front of mind from the very beginning.

The political narrative that Bannon spun went like this: Coastal elites who are politically and culturally disconnected from “ordinary Americans” have pried open America’s borders to foreign-made commodities, international capital and culturally alien peoples. While these “globalist” policies have enriched the elites, they’ve also led to a deteriorating standard of living for working-class Americans. The influx of foreigners (who look different, sound different and even worship differently) has irrevocably altered the fabric of American life, says Bannon.

As with any good piece of propaganda, there is some truth in the storyline. The process we call “globalization” has been authored by the very elites who have been its principle beneficiaries. It is also true that open borders for commodities and capital tend to be married to liberal immigration policies. As a consequence, in recent decades international trade and investment have increased along with higher levels of immigration, especially that from non-English-speaking countries.

This process of opening borders to goods, investment and peoples has unfolded side-by-side with deindustrialization, wage stagnation and greater income inequality.

Bannon’s politics have all the markers of propagandistic coherence, too: He pinpoints a problem (open borders), a sacrificial victim (“ordinary” Americans), an identifiable scapegoat (coastal elites, foreigners), and suggests a solution to the problem (“build the wall!”). The effectiveness of the message resides in its simplicity.

But is the narrative factually correct?

While it may be soothing for more liberally minded people to ascribe mass acceptance of Bannon’s narrative to “economic illiteracy,” it remains the case that this narrative has been successful. Is the electoral potency of Trumpism predicated on a catalogue of economic errors?

Conventional narrative about globalization and deindustrialization

The conventional storyline about the decline of manufacturing in advanced economies tends to emphasize technological factors, including automation, rather than public policy choices.

The dual forces of automation and offshoring have reduced the demand for Canadian and American manufacturing workers, we are told, which has in turn contributed to the stagnation of manufacturing wages. In the same way that manufacturing displaced agriculture in the early part of the 20th century, the knowledge economy has dislodged manufacturing.

The declining significance of manufacturing can be seen in figure 1, which chronicles the manufacturing share of overall employment in Canada (red line) and the United States (blue line). In the mid-1960s, roughly one-quarter of Canadian and American workers were employed at a workplace that manufactured a commodity. Manufacturing steadily declined in both countries such that, at present, fewer than 1 in 10 workers are employed in manufacturing. The picture does not materially alter if, instead of manufacturing employment, we substitute the share of business investment. There, too, we see a steady decline in the economic significance of manufacturing over the past half century.

These trend lines seem to undermine the Bannon-Trump narrative that manufacturing has been traded away by globalist elites. If technological change has fuelled the decline of formerly omnipotent industries, public policy responses are likely to be ineffective. However, acceptance of the “inevitability” of manufacturing’s decline lacks nuance.

Some diversion from the conventional storyline

Consider, for starters, the fact that some other advanced economies have done a much better job retaining their manufacturing bases. Using value-added as the metric, manufacturing makes up 23 percent of German output and 21 percent of Japanese output, in comparison with just 12 percent in the United States and 10 percent in Canada. According to the World Development Indicators the economic significance of Germany’s manufacturing sector has remained unchanged since the mid-1990s, while Japan saw a slight decrease in manufacturing’s share of total output, from 24 percent in the mid-1990s.

Also consider figure 2, which plots the relative share of high-tech manufacturing investment in Canada and the United States. The red line, on the left axis, captures fixed-asset investment (in nonresidential structures, machinery and equipment) in Canadian transportation equipment manufacturing (including auto, aerospace and rail), while the blue line, on the right axis, captures fixed-asset investment in US motor vehicle manufacturing.

Both are plotted as a share of total private fixed-asset investment. (Canada lacks auto investment information over this time period, making an identical comparison impossible.) The picture that emerges from figure 2 differs markedly from that in figure 1. Instead of a steady decline over the past half century, we see a nonlinear trend: In the three decades to in 1994 when NAFTA took effect, advanced manufacturing investment trended upward; in the two decades since NAFTA, the trend line is unambiguously negative.

If technological change was the main driver of deindustrialization, then we should have seen a comparable decline in the manufacturing share of output across all OECD countries, but that’s not what the data tell us. These facts seem to upend the claim that automation is driving the decline of Canadian and American manufacturing. These facts also fly in the face of the claims made by some commentators that NAFTA helped preserve, rather than undermine, US auto manufacturing.

And there is even more to support Bannon’s storyline (one that Unifor and other labour unions have been arguing for more than a decade), that the erosion of America’s industrial base helped Trump win in 2016. The data suggest that a few key states handed the election to Trump: Michigan, Ohio, Wisconsin and Pennsylvania. These rustbelt states traditionally voted Democrat, but swung Republican in 2016. A study conducted in 2017 found that a significant proportion of the white working class who voted for Obama in 2012 but pivoted toward Trump in 2016 felt that they were struggling to keep up with the rising cost of living. Many reported feeling abandoned by the Democratic Party’s economic policy. Trump’s message of restoring America’s battered industrial base appears to have resonated. Is there any truth to these beliefs?

Figure 3 plots the average hourly earnings in Canadian and American manufacturing (inflation-adjusted) over the past seven decades. The blue line captures the American experience and clearly demonstrates deteriorating wages for US manufacturing workers. Between 1978 and 2016, American manufacturing wages fell by 8 percent. Working-class wage suppression and the associated deterioration of blue-collar living standards is real.

The Canadian story is more complicated. Despite a significant increase in labour productivity over the past four decades, Canadian manufacturing wage levels are only 14 percent higher today than they were in 1978. Since NAFTA came into effect in 1994, Canada’s manufacturing workers have barely seen any improvement in their standard of living, registering just a 9 percent wage increase.

Canada’s manufacturing workers are in a similar situation to that of their American counterparts: Job destruction has been on the rise, while wage levels have either stagnated or fallen. One reason why Canadian manufacturing workers have fared better than American workers may be the higher rate of unionization. Fewer than 1 in 10 manufacturing workers in the US are represented by unions, while in Canada the figure is 1 in 4. Canadian manufacturing workers therefore have much more bargaining power than their American counterparts, which helps them secure higher wages (as well as greater public policy clout).

Mexico and the North American auto industry

So far the data seems to support key claims made by Bannon and Trump (claims that have been long advanced by the social democratic left, Unifor included) — the coming into force of NAFTA corresponds with the deterioration of America’s advanced manufacturing base, and blue-collar workers have struggled to keep up with the cost of living. Trump has targeted Mexico, saying that it “stole” jobs from Americans. But is that really true? After all, you can’t bring the jobs back if they were never exported in the first place.

Consider figure 4, which indexes North American auto production (to 100 in 1994 in order to capture the relative change) among the three NAFTA partners. The red and blue lines capture the Canadian and American experience, while the green line captures Mexico’s finished vehicle output. Comparing 2016 with 1994, it turns out that Canadian auto production has only increased by 2 percent, while American motor vehicle output has contracted by 1 percent. Mexico’s auto production has soared — more than tripling in the past 22 years.

Now consider figure 5, which registers the cumulative change in the share of North American auto production of the three NAFTA partners. In the first decade of NAFTA’s existence, the growth of auto production was balanced, with relative gains and losses among the NAFTA partners amounting to a net change in their production share of just 3 percent. After 2004, Mexico captured the entirety of the gains in North American production, at the expense of both Canada and the United States. By 2016, Canada had lost three percent of its production share, the US had shed 9 percent and Mexico had captured 12 percent. Mexico surpassed Canada in 2008 in terms of finished vehicle output, and by 2016 it accounted for 20 percent of North American production. Canada now accounts for just 13 percent of NAFTA-zone finished vehicle output.

Auto production among the three NAFTA partners has increased by 16 percent since 1994, rising from 15.7 million units to 18.1 million units. In the context of growing demand for auto manufacturing inputs and outputs, there’s no reason Canada would not capture a portion of that growth. However, in 2016, Canada only assembled 49,000 more units than it did in 1994, while the United States assembled 61,000 fewer units. This means that Mexico has captured the entire increase in North American auto production since NAFTA came into effect. Indeed, Mexico captured slightly more than 100 percent of the net increase due to the net decline in the United States.

These figures capture the change in final vehicle assembly, ignoring the even more lopsided distribution of overall auto employment, including auto parts. Despite the fact that auto production and consumption have risen, Canadian auto employment has contracted. Since 2004, when Mexico auto production began to take off, Canada shed some 38,000 auto jobs (nearly one in four positions). Canadian employment in the Detroit Three producers — GM, Ford, Fiat-Chrysler — has halved since NAFTA came into effect, falling from 52,000 in 1993 to 23,000 by 2016. The United States hasn’t fared much better, losing 173,000 — 1 in 6 — auto jobs since 2004. American auto employment is down even further since 1994, shedding some 228,000 positions, or one in five.

In the past five years, 11 new auto plants have been built in North America: 9  in Mexico and 2 in the southern United States. Both Canada and the United States run significant auto trade deficits with Mexico. As Unifor and others have long argued, Mexico has in fact benefitted enormously at the expense of Canadian and American workers due to sharp wage and regulatory differences among the NAFTA countries.

Canadian auto workers have long understood that their jobs are at risk of being shipped to Mexico, up to and including recent announcements that the GMC Terrain and Toyota Corolla, made in the Ontario communities of Ingersoll and Cambridge, respectively, will soon be assembled in Mexico. However, because Mexican auto workers cannot afford to buy the cars they build, these vehicles will be shipped back to Canadian and American consumers duty-free. While Mexico accounts for 45 percent of North American auto jobs, it only accounts for 8 percent of North American auto sales (as detailed in Unifor’s positioning statement on autos and NAFTA).

Trump’s claim that the Mexicans are “stealing our jobs,” therefore, contains some truth. The message is sloppily and opportunistically crafted, and it unfairly scapegoats Mexican workers themselves, but for many blue-collar workers in the United States and Canada, it has an emotional impact and is politically salient. Using virtually any metric — investment, finished vehicle production, employment, trade flows — the result is the same: The North American auto industry has grown, but it has migrated south of the US border, taking many jobs with it.

Economic competition, the doctrine of free trade and Mexican wages 

How have the Mexicans people fared under NAFTA? A significant expansion in the demand for labour as a result of the arrival of a high-productivity industry should, if we are to believe conventional economic thinking, lead to an increase in wages and benefits. In classical economic theory, liberalized trade produces net benefits for all trading partners, even in the extreme instance where one trading partner has an absolute advantage in the production of the traded commodities. Paul Samuelson’s factor price equalization theory (which has held considerable sway in the field of economics) predicts that the market price of identical factors of production — the manufacturing wage rate among trading auto jurisdictions, for example — should converge with the onset of open trade. In other words, Mexican wages should have risen.

From an international manufacturing standpoint, hypothetically NAFTA could be defended if the economic decline experienced by Canadian and American workers was offset by improvements in the standard of living of Mexican workers. Does the data support this interpretation?

Figure 6 shows the cumulative change in Mexico’s share of North American finished vehicle output and in its manufacturing-sector wage levels. Even as auto investment poured into Mexico after 2004, manufacturing wages in Mexico declined, having fallen by 27 percent between 2002 and 2016. The relationship between a strong and growing manufacturing base (25 percent of Mexicans work in manufacturing today, up from 21 percent in 1993) and the deteriorating wages of manufacturing workers is sharply at odds with the central tenets of mainstream economic thinking.

Why have Mexican wages failed to rise? The fact that Mexico mostly lacks independent labour unions is undoubtedly one factor. Many Mexican auto workers who fall under so-called “protection contracts” do not even know if they belong to a union, let alone being able to actively bargain with their employer. The Mexican state and the auto companies effectively determine wage levels, with little or no input from the workers.

Exhibit A in this tale is a 2014 sweetheart agreement signed between BMW and Mexico’s union officials (note that in Mexico, state unions are answerable to the government and so lack independence) — an agreement that was inked before the workers were even hired. On the basis of an agreement that had absolutely no input from the workers themselves, assembly-line workers in San Luis Potosi will build high-end vehicles for just US$1.10 per hour to start, going up to only US$2.53 per hour, according to a Bloomberg report. That’s in line with the average wage rate of US$2.04 in Mexico’s auto plants.

When the Mexican state itself is in the wage suppression game, there is very little Canadian or American auto workers can do to make themselves more “cost-competitive” with Mexico, despite the high quality of the products they build.

While it’s almost certainly an understatement to say that the US president has an uncomfortably distant relationship with the truth, it would be a serious mistake to dismiss as untrue everything Trump says. On the issue of economic development, there is some factual support for aspects of his message. Manufacturing industries in Canada and the United States have declined considerably since the implementation of NAFTA. This is despite the fact that the demand for automobiles and auto parts has grown. Auto workers have seen their jobs shipped to low-wage and weakly regulated jurisdictions, including Mexico. This industrial migration erodes blue-collar living standards.

The facts support the existence of links between liberalized trade, deindustrialization and working-class wage suppression. One thing that the election of Donald Trump demonstrates is that the liberal international order, including economic integration, can be reversed. In a democratic political context, international trade agreements require mass political support. They are not the inevitable result of technological change. And free-trade agreements will not be supported if they threaten people’s livelihoods.

Photo: Shutterstock, by  StonePhotos.

Do you have something to say about the article you just read? Be part of the Policy Options discussion, and send in your own submission. Here is a link on how to do it. | Souhaitez-vous réagir à cet article ? Joignez-vous aux débats d’Options politiques et soumettez-nous votre texte en suivant ces directives.

Jordan Brennan
Jordan Brennan is an unrepentant Torontonian, an economist with Unifor and a visiting scholar at Harvard Law School.

You are welcome to republish this Policy Options article online or in print periodicals, under a Creative Commons/No Derivatives licence.

Creative Commons License

More like this