Commerce, jobs and politics: the impact of the USA–China trade on USA domestic politics

Abstract

Purpose

The purpose of the paper “Commerce, jobs and politics: the impact of the USA–China trade on USA domestic politics” is to examine the impact of Chinese trade with the USA to determine the consequences of the trade on manufacturing employment. The geographic and sectoral impacts of this trade are assessed. The conclusion is that the USA–China trade has affected political polarization in such a way as to affect electoral outcomes. Implications for policy are discussed in the paper.

Design/methodology/approach

The overall design is a focused case study in terms of its focus on the USA–China trade relations. There is also a statistical component due to the breakdown of the USA in economic commuting zones.

Findings

The major finding is that Chinese import penetration created substantial political polarization in the USA and that polarization affected electoral outcomes. Chinese import penetration also resulted in a shift of jobs from the eastern heartland to the coasts. Much of the transition was aided by the restructuring of jobs within firms from manufacturing to high-end services.

Research limitations/implications

Perhaps, the biggest limitation concerns how general and durable the findings are. The authors establish that the first decade after Chinese entry into the World Trade Organization (WTO) (2001) was characterized by economic disruption in the USA labor market. Whether the economic effects will have a longer duration is not known.

Practical implications

One practical limitation is that it is difficult to know what policy actions to take on the basis of the research: trade policy, human capital (education) policy or place-based policies which aid particular regions.

Social implications

The social implications in this paper are jobs and employment policy.

Originality/value

The author thinks this is very original work, though based on the work of several economists. But outside of a few articles, the author does not think much has appeared in political science journals.

Citation

Caporaso, J. (2022), "Commerce, jobs and politics: the impact of the USA–China trade on USA domestic politics", International Trade, Politics and Development, Vol. 6 No. 1, pp. 2-13. https://doi.org/10.1108/ITPD-10-2021-0013

1. Introduction

Because of its dual focus on economics and politics, International Trade, Politics and Development provides a vehicle for understanding the political economy of international trade. This is fortunate because the subject of this paper, trade relations between the USA and China has generated enormous political controversy from the adjustment of trade frictions related to tariffs and intellectual property rights all the way to the possibility of weaponized interdependence and war. Titles of recent books such as Unbalanced (Roach, 2014), The Long Game (Doshi, 2021) and The World Turned Upside Down (Prestowitz, 2021) provide a flavor of this highly contentious trade relationship. In this article, we assess the political and economic significance of the Chinese–USA trade relationship.

Our focus is on the consequences of China–USA trade for USA jobs and electoral politics. The job effects of trade are difficult to untangle from a host of other considerations, notably automation and shifting comparative advantage. What can be ascribed to a particular country is even trickier, since bilateral trade relations are notoriously sensitive to changes among numerous trade partners. If workers in Country A are losing jobs due to trade with Country B, putting tariffs in place may simply shift trade to Country C. Also, the causal chain from trade to elections is a long one, with many links, each of which has to be empirically established.

While we can't explore all causes of trade frictions between the USA and China, we focus on trade with China associated with entry into the WTO in 2001 and make the case that there was a substantial job loss effect in the USA. While job losses had a limited temporal duration (Bloom et al., 2019), the political repercussions endure until today and have become part of the standard grievance narrative that fuels support for right wing of the Republican Party. Thus, our focus goes beyond the gains from trade to take into account the restructuring of economic activity, job losses and gains, political mobilization and votes [1].

We proceed as follows: First, we describe the USA–China trade imbalance. Second, we assess whether China's trade is responsible for USA job losses. Third, we examine the geographic distribution of job losses. Fourth, we examine the political implications of job losses. Because job losses are geographically concentrated, they have become politically significant. The loss of three million jobs, just to take a figure, interspersed randomly across a vast country of 330 million people and presents a more difficult situation for consciousness raising, party mobilization and policy making than one in which the damages are concentrated in one area.

2. The USA–China trade imbalance

Trade between the USA and China has been unbalanced for well over 20 years. However, once China gained membership in the WTO (2001) Chinese exports to the USA increased faster than USA exports to China. China had a trade surplus of US$80bn with the USA in 2001. Charlene Barshevsky, United States Trade Representative (USTR) at the time argued that the USA trade deficit would be cut dramatically once China acquired membership (Prestowitz, 2021, p. 3). Instead, the USA current account deficit with China ballooned to US$400bn by 2018.

USA multinational companies favored Chinese access to the WTO and with the decision to grant China the much sought-after permanent normalized trade relations (PNTR) status. Before PNTR, China's most favored nation (MFN) status had to be approved on a yearly basis by a vote of the USA Congress. The USA multinational corporations were constrained since establishing plants, factories and managerial facilities in China involved long-term planning and sunk costs. Firms did not want to move if, once there, MFN status could be revoked by a simple act of the USA Congress. After PNTRs was established, multinationals were less vulnerable to short-term political pressures. This allowed USA companies to increase their production outsourcing to China to make products many of which were then exported to the USA.

China's trade surplus allowed it to finance a manufacturing boom, to increase dramatically exports and to invest heavily in USA treasuries as payment for their surplus. Their investments in USA treasuries, while yielding a low return, allowed investment in productive enterprises within China. At the same time, the USA could maintain its high rate of consumption and a low interest rate environment, which, in turn, helped to support a growing housing market. Chinese capital flows to the USA, the mirror image of the USA trade deficit, made it easier for businesses to invest and for prospective home-owners to qualify for mortgages.

3. USA job losses associated with the trade imbalance

The estimated size of the USA job losses in the manufacturing sector due to trade with China varies widely. One of the most authoritative econometric studies by David Autor et al. “The China Syndrome,” (2013) judges that there were substantial impacts on employment (2–3 million), labor force participation and wages due to import competition during the period of their analysis, 1990 to 2007. Case and Deaton, in their book, Deaths of Despair (2020, p. 219), are persuaded by the research of Autor et al. and accept the 2–3 million lost jobs due to China's accession to the WTO. Case and Deaton's work goes well beyond trade imbalances and employment and seeks to describe the human costs of these job losses in terms of status of the unemployed, destabilization of families and increase of morbidity and mortality, especially among white males, who held many of these privileged manufacturing positions.

The circumstantial evidence of a China effect on the USA trade and jobs is strong (Autor et al., 2013, Pierce and Schott, 2016 and Acemoglu et al., 2016). By 2000, the low-income country share of the USA imports was 15% and this escalated rapidly so that by 2007 it was 28%. Roughly 89% of this incremental growth was due to growth in Chinese imports. Increases in the USA spending on Chinese imports followed this pattern of import growth, and while this is part of a longer-term trend that may lead some to discount the effect of WTO membership, there is a noticeable inflection point in the USA consumption of Chinese goods around 2001. During the same period (1990–2007), the share of the USA workforce in manufacturing declined precipitously from 12.6% to 8.4 (Autor et al., 2013, p. 2122).

Further, the employment effect of trade with China was abrupt. Pierce and Schott (2016) refer to a “surprisingly swift decline in US manufacturing employment,” while Acemoglu et al. (2016) refer to the “Great US Employment Sag of the 2000s” Pierce and Schott (2016) carefully document a dramatic decline in the USA manufacturing employment in the period 2000–2007, exactly the period of WTO membership for China (2001) and the granting of PNTR status by the USA Congress to China (October 2000), which became effective when China entered the WTO.

The change in the status of trade relations was and continues to be of great importance for China and the USA since it lowered uncertainty. Previously, most favored nation status was granted to China annually upon a vote of the Congress. Now MFN status became “normal.” The uncertainty affected the USA more than China. Trade planning could be done on an annual basis. Investments by multinational firms, however, since they involve fixed costs and long time-horizons, could greatly benefit from advance planning. The USA corporations, particularly large multinational corporations, were very enthusiastic about the liberalization of trade with China. The USA–China Business Council and the Business Roundtable lobbied in this cause (Prestovitz, 2021, pp. 200–206).

We should point out that consumption of Chinese imported goods as well as declining the USA employment in the manufacturing sector are part of longer term secular trends. Automation, trade with China and imports of manufactured goods from multiple countries all contributed to competition within the manufacturing sectors. However, a close look at the employment figures in the USA provides evidence of a China effect [2].

Finally, the striking way in which China's surplus mirrored the USA deficits (roughly 5% surplus for China and roughly the same amount on the deficit side for the USA) fueled popular opinion that the USA was China's “demander of last resort” (Aizenman and Jinjarak, 2008). This perception prevailed despite the fact that a 5% surplus for China and 5% deficit for the USA could result from countless different combinations of bilateral trade balances. In this sense, focusing too much on trade imbalances with a single country is misleading, since if the USA decreased imports from China without changing its savings rate, it would “only” reallocate China's surpluses to other trade partners. This is where a purely economic approach to trade misses a central point, namely that with whom one has trade imbalances matters as much, though in different ways, as the overall trade balance. The scholarly discussion that bilateral trade balances are misleading, because they are easily shifted from one country to another, is true but somewhat beside the point [3]. From a comprehensive political economy viewpoint, one that centers power as well as wealth (Hirschman, 1945 and Keohane, 1984), there may be good reason to fear concentrating deficits in one trade partner.

4. Geographic distribution of job losses

We start from three observations. First, there have been losses of manufacturing jobs in the USA due to trade with China. Second, these losses have a distinctive geographic imprint. For the most part, trade with China has reallocated jobs from the heartland (eastern heartland) and the south to the coastal areas (Austin et al., 2018) at the same time shifting the composition of the USA exports from manufacturing to services. The regional and sectoral composition of trade has shifted at the same time and sometimes with respect to the same regions, making it difficult to disentangle the overall effects (Bloom et al., 2019). Third, the effects of this economic restructuring have started to register in the mobilization of grievances along geographic, gender and racial lines (Autor et al., 2020, pp. 3145–3152).

The first proposition, concerning job losses due to the China trade, is well established (Autor et al., 2016, Acemoglu et al., 2016 and Pierce and Schott, 2016). The second proposition, concerning the geographic effect of trade on jobs, is well accepted at the general level – the mid-west, or heartland, or rust-belt has been hurt, the coasts have gained – but there is considerable disagreement on important issues, most importantly how to divide the USA geographically. What exactly is the mid-west? Is it different from the two heartlands discussed by Austin et al. (2018). The third proposition, concerning the mobilization of these cleavages, is not well understood. Here we take advantage of some recent research in political science and economics, particularly Autor et al.'s “Importing Political Polarization” (2020).

4.1 Geographic and sectoral profiles of employment changes

A cautious analysis of changing job patterns in the mid-west and south shows that at the margins, manufacturing jobs were lost without replacement. This claim requires that we disaggregate the “mid-west” into what Austin et al. call the “eastern and western heartlands” since these two parts of the country have very different profiles, with the western heartlands excelling in growth, while the eastern heartland is the slowest growing region of the country from 1965 to 2015 (Austin et al., 2018, p. 152). The coastal areas also experienced job losses in manufacturing but these losses were offset by new jobs in the service sector (Feenstra et al., 2018), often service jobs within manufacturing firms, so-called “white-collar manufacturers”. Bloom et al. (2019, p. 1) argue that manufacturing job losses were offset by service jobs in the coastal areas and they interpret these patterns as part of a single process of restructuring by multinational firms. Trade with China reallocated jobs from the interior of the USA (excluding the western heartland, from the Dakotas to Texas north to south) to the coasts and from manufacturing to services (Bloom et al., 2019, p. 1).

4.2 Job losses in the eastern heartland and south

Growth in income and wealth in the USA has gone hand in hand with job losses. As Austin et al. show (2018, p. 164) as of 2015, roughly 10% of working age males were not in the workforce (not working and not looking for work). This is in addition to the unemployment rate regularly reported by the Bureau of Labor Statistics. Moretti, an economic geographer, calls this the “great divergence”. According to Moretti, “Technological change and globalization result in more employment opportunities for a low-skilled worker in a high-tech hub but fewer opportunities for a similar worker in a hollowed-out manufacturing town.” There is a huge difference between Seattle and Boston on the one hand, and Akron, Ohio and Flint, Michigan on the other.

Perhaps, the strongest and most widely accepted claim is that China trade involves manufacturing jobs lost in the heartland (see Austin et al., 2018 for helpful spatial refinements) along with job gains in the coastal areas. This is not because coastal areas avoided economic dislocation and pain. Indeed, there were manufacturing job losses in the coastal areas too but overall restructuring involved the addition of jobs in the service sectors. Since many of these service sector jobs were high-quality positions (management, research and development and design), this softened the blow of jobs lost in manufacturing (Bloom, 2019, pp. 2–3).

In order to make sense of shifting job patterns, we have to rethink traditional geographic categories. In particular, the geographic categories coastal states, mid-west and south are not very useful partly because the coastal areas are heterogeneous but also because the western heartland is doing well, while other parts (the eastern parts) suffer. The heterogeneity of the coastal states is not between east and west coasts but within each coastal region, i.e. there is a huge difference between New York City and upstate New York and between San Francisco and “inland” cities such as Stockton, Yuba City and Modesto, California. The geographic classification developed by Austin et al. (2018) is more useful for our purposes. The authors divide the USA into prosperous coastal areas, western heartland and eastern heartland. The south, usually thought of as a distinct region, is treated either as part of the eastern heartland or the southeastern coastal area.

Within this framework, we can evaluate the claim that the coasts gained and the heartland lost in a more nuanced fashion. The first thing to note is that the western heartland is not suffering economically and did not sustain the heavy manufacturing job losses related to China that affected other parts of the country. The western heartland runs north to south from North Dakota to Texas (North Dakota, South Dakota, Nebraska, Kansas, Oklahoma and Texas). This vertical region is characterized by relatively efficient export-oriented agriculture, food processing, farm equipment and ranching, while the southern part, Texas, thrives not only as a producer of primary products (beef and oil), but also as a manufacturing area centering on computer components and the aeronautics industry. Texas is the number one exporting state in the USA.

It is a mistake to include the western heartland states along with the vast interior of the USA that is non-coastal; because by doing, so we aggregate quite different social and economic phenomena. If manufacturing job losses due to trade are a cause of political polarization in the USA, we have to identify geographical areas within the country that are high on this dimension and separate them from those that are low. The geographic classification of Austin et al. (2018) is a step in the right direction.

The areas of the western heartland are doing well in terms of employment, efficiency and export orientation and do not suffer from the malaise and health problems (morbidity, longevity and the opioid epidemic) of the eastern heartland states.

The eastern heartland presents a different picture. The eastern heartland refers to the areas east of the Mississippi but not on the coasts. Again, we can trace a line north to south from Michigan and Wisconsin to Mississippi. This line would go through Wisconsin, Michigan, Indiana, Ohio, Kentucky, Tennessee, Mississippi and Alabama. Some contiguous states, e.g. Pennsylvania, display an economic profile of declining industrial or commercial towns.

Austin et al. report that growth in the real economies of these three regions was as follows, 1965 to 2016: coastal areas grew by 342%, western heartland grew by 475% and eastern heartland by 187% (Austin et al., 2018, p. 170). These figures will surprise many people who mistakenly think of the western interior of the USA as isolated, agricultural (it is, but very efficient) and cut off from the human capital externalities and agglomeration economies of the coastal areas. But what do the economic conditions of the heartland tell us about trade with China in manufacturing? We return to this question after surveying conditions on the coasts so that we can treat all regions together.

4.3 Job gains on the coasts

There are two lessons to be drawn from the picture of jobs on the coasts. The first is that there were manufacturing job losses on the coasts but that these losses came with compensation from two sources: first, the addition of new jobs in manufacturing because of increasing exports, and second, the increase in substantial numbers of well-paying service jobs. The addition of these medium and high-tech service positions, along with the regional externalities (learning externalities and network effects) that are caused by the presence of dense concentrations of human capital, softened the impact of manufacturing job losses in some parts of the coastal areas. In the next section, we examine the interaction between the sectoral and geographic aspects of economic restructuring.

4.4 Sectoral shifts and job restructuring

Manufacturing jobs lost in the eastern heartland and the south were not compensated by shifts in the service sectors of the economy. Bloom et al. argue that trade with China resulted in a double reallocation from manufacturing to services and from the heartland to the coasts (2019, p. 1). They do not say which heartland, since they do not follow the Austin et al. regional categories. In the coastal areas such compensation took place mostly through market mechanisms, though of course the USA Government played a major role in conferring most favored nation status on China on a permanent basis.

Often, the addition of service jobs was an integral part of the same processes of plant closures and downsizing that took place in manufacturing industries. This runs counter to the usual pattern in which a loss of manufacturing jobs entails collateral losses in services. Moretti's research shows that for every manufacturing job lost, there are 1.6 additional jobs lost in services such as barbers, waitresses, doctors, builders and retailers (2013, p. 24). Here we find a contradictory trend in which creation of service jobs seem to be part of the same process in which manufacturing jobs are destroyed.

To see this better, we can partition changes in employment associated with the China trade into three categories: job losses due to firm failure and closure; job losses due to downsizing and plant (but not firm) closure and jobs that have been rebranded, or reincarnated as it were, from manufacturing to service (Bloom et al., 2019, pp. 1–2 and abstract). This third category turns out to be very important. The reorganization of production takes place within the same firms. We know this since very few firms died out and the ones which “shed” manufacturing jobs were often the ones which rebranded substantial parts of their workforce as service. Bloom et al., on the basis of extensive firm-level research, state flatly that “Plant closures from firm death do not play a major role in driving manufacturing job losses”. (2019, p. 12) Rather, the authors conclude that “the net switching to services in high human capital areas accounts for over 50% of the total negative effect on manufacturing jobs”. (2019, p. 15) The effects in low human capital areas were much smaller and not (statistically) significant.

Thus, “the effect of Chinese imports on establishments switching from manufacturing to related services discussed above is concentrated primarily in high human capital areas” (Bloom et al., 2019, p. 15). The transition is from manufacturing to service jobs, since it exerts its most powerful leverage within areas that are high in human capital, and since these areas are already the richest and most dynamic parts of the USA economy (see Moretti, 2013, pp. 94–95), they inevitably produce greater inequality among regions. This protection harms poorer individuals (on the consumption side) while at the same time protects firms with influential lobbies. The Schumpeterian process of creative destruction is at work, destroying old jobs and creating new ones but the changes are from manufacturing to services rather than from less to more sophisticated manufacturing. The net welfare effects may be positive or negative, but the changes are more harmful to less educated manual workers.

5. Trade, jobs, polarization and USA politics

This article attempts to integrate three large pieces of a puzzle: one, the USA–China trade itself; two, the damage done to USA jobs by this trade and three, the mobilization of these losses into grievances. Only the first piece is straightforward. The second piece requires establishing a connection between trade and USA job losses. The third piece is the most difficult. Once job losses are established, we need to show the transition from objective job losses to collective awareness, mobilization and political action. Such a transition involves a large collective action problem, since the damage involves millions of workers located in different firms and different places.

The rest of this paper focuses on the remaining but all-important piece of the puzzle, the political implications of trade. This piece is difficult to establish conclusively for several reasons. The key limitation is that we have no overall model of the political process to rival economic models of how trade affects jobs and incomes [4]. How, or even if, economic cleavages are translated into social and political movements is not a straightforward matter.

To make the case that trade affects political mobilization and electoral outcomes, we start with the Stolper–Samuelson theorem (1941) and with Ronald Rogowski's (1989) effort to extend this theorem to political settings. We then link these two lines of research to the work of Autor et al. (2013, 2016 and 2020) in their attempt to show that trade not only causes economic cleavages along predictable lines, but also that these “objective” economic cleavages (income differences and jobs) are being transformed into active political forces. In short, those whose incomes have been affected by trade, whether in a positive or negative direction, are mobilizing in a variety of ways to advance their positions.

In 1941, Wolfgang Stolpler and Paul Samuelson (1941) penned a remarkable article ripe with implications for scholarship in both economics and political science. The basic argument of “Protection and Real Wages” had to do with changing incomes of holders of different factors of production as a function of international trade. What they attempted to show is that free trade harms, in relative terms, those owners of factors (labor, capital and land) that are relatively scarce and benefits those who are holders of relatively abundant factors. Simply put, free trade will benefit workers if labor is the relatively abundant factor, capitalists if capital is the relative abundant factor and farmers if land is well endowed relative to trade partners [5]. To take a concrete example, if a free trade agreement such as the North American Free Trade Agreement were to bring together a resource-rich country (Canada), a labor rich country (Mexico) and a capital rich country (the USA), the prediction would be that landowners, workers and capitalists would gain respectively in Canada, Mexico and the USA.

Despite the rich implications of the Stolper–Samuelson theorem for politics, it was largely unrecognized in the political science literature [6]. Yet, economic cleavages caused by trade provide raw material for political analysis. This period of dormancy has been partly corrected by Ron Rogowski's Commerce and Coalitions (1989). Rogowski argues that the Stolper–Samuelson theorem need not stop with trade's impact on relative incomes. Changes in income are unlikely to be accepted passively by both winners or losers. Those who gain (lose) are likely to recognize such and to take action to further (resist) these trends.

Rogowski (1989, pp. 4–5) makes three reasonable assumptions to link economic cleavages and political mobilization. First, that those who gain or lose from trade will recognize it and attempt to further or retard these change. Second, those who increase wealth will expand political influence, a proposition supported by much research since Commerce and Coalitions (Gilens, 2014 and Bartels, 2008). Third, political entrepreneurs will recognize the opportunity of exploiting these cleavages for political gain and provide political channels to convert demands into public actions and programs.

Rogowski provides a vantage point for viewing the political implications of trade. He makes a clear theoretical case for the links between trade-induced cleavages and politics. However, for his insights to be convincing, we need to establish the connections with hard empirical evidence. Inequalities in the distribution of income, whether caused by trade or some other factor, may remain just that, inequalities. Indeed, considerable research on inequality (Gilens, 2014 and Bartels, 2008) more or less supports this conclusion. Some objective inequalities call out for political reform, while others lie dormant. Schattschneider, in The Semi-Sovereign People (1960), argues that only some objective interests are mobilized. For a political problem to be seen as a political problem, it has to be “named,” “blamed” and “claimed”. (Felstiner et al., 1981, pp. 631–637).

Due to recent research, particularly that of Autor et al. (2020), we now have some leverage over questions that have eluded us for a long time. This research is reported in “Importing Political Polarization: the Electoral Consequences of Rising Trade Exposure.” (2020). Autor et al. (2020) avoid using large spatial units such as areas of the country (mid-west and east coast) and even states. As has been pointed out, many states are internally very heterogeneous with some parts trade exposed and others relatively isolated. States turn out to have center-periphery characteristics that mirror the country as a whole. Instead, the authors construct smaller and economically tractable units which they call “commuting zones,” defined as clusters of adjacent counties which form labor markets and areas within which workers commute to work. (2020, p. 3152) These zones may combine or divide counties in diverse ways. The task is to match them with political districts. For each commuting zone, they measure the China shock as the average change in import penetration due to Chinese imports over the industries of that area weighted by the share of each industry in the commuting zone's initial employment. The authors then align the units of economic activity and labor, on the one hand, and with units of political representation, on the other (see 2020, pp. 3141-3142).

5.1 Increasing polarization of news consumption

Ideological polarization is measured at the commuting zone level as the proportion of people watching Fox News, MSNBC (Microsoft and National Broadcasting Corporation) and CNN (Cable News Network). The results show clearly that the greater the import penetration, the greater the shift to Fox News by a substantial margin. And Fox viewers faithfully vote for the Republican candidate by high margins. In the 2004 and 2016 elections, Fox News viewers voted for the Republican candidate by margins of 62 and 66% (2020, p. 3158). The increase in Fox News viewers was mirrored by the losses to viewership of CNN and MSNBC. There is no comparable shift to the left the way conservatives move to the right. However, the important message remains that trade exposure to China, by commuting zone, “moves the ideological needle of media consumption rightward among non-minority households (2020, p. 3161).”

5.2 Increasing polarization of campaign contributions

The second aspect of the importation of polarization concerns the pattern of campaign contributions. Voters naturally give donations to parties and candidates of their choice. Individuals and corporations can give to both parties at once or increase and decrease contributions in line with shifting divides regarding issue salience and identity. Contributions from both the left and the right increased to their favored party, but contributions from the middle of the political spectrum did not. Polarization may not have increased in terms of the party distance of those making political donations but it has in terms of clustering and emptying out of the middle (Autor et al., 2020, p. 3174).

The relationship between political attitudes of people in commuting zones and campaign contributions is central. Partisan “giving” is not evidence per se that donations are a reflection of trade with China. However, the authors have identified a close link, an unmistakable one, between increases in trade exposure with China and partisan donations. The first piece of evidence is that districts with larger increases in China trade have larger increases in campaign contributions (2020, p. 3162). As to the ideological composition of the electorate with regard to campaign contributions, greater trade exposure increases giving by both the left and the right and is substantially higher on the left. (2020, p. 3163) Once again, giving by the moderates, the ideological middle declines. This pattern of increased campaign contributions by left and right, with a corollary decrease in contributions by moderates, is sustained from 2000 to 2016. And this pattern is most pronounced in the most trade exposed districts (2020, p. 3165).

5.3 Polarization and voting

The third and most important political aspect of the USA–China trade concerns voting in Congressional and Presidential elections. Recall that trade exposure to China in manufactures has led to declining jobs and incomes, increasing non-labor force participation and disruptive social consequences depicted in Deaths of Despair. (2020) It has also led to a decline in the attractiveness of prime working age men as marriage partners, which was vividly described by Autor et al. (2019) in “When Work Disappears.” The damage is clear but how, if at all, does this economic and social damage affect voting patterns? Looking at Congressional elections first, as trade exposure to China increases in local labor markets and voting districts, extreme voting increases in both parties. The authors debunk the claim that there is only a rightward shift with no parallel on the left. Yet, it is the Republican Party which in the net gains seats. This is where the disaggregated units prove of value. Candidates of the right pick up their votes in highly competitive districts and are able to flip them to their party. Democratic candidates increase votes and, indeed, vote share in less-competitive districts. However, since the USA electoral rules are based on single member districts, winner-take-all and plurality vote, increasing one's share from 40 to 45% of the vote does not matter if there are only two candidates. Extremist voting is at the expense of moderate Democrats in both cases.

In presidential voting, the patterns are similar though the effects are smaller, and votes are aggregated at the state level and allocated with few exceptions (Maine and Nebraska) on a winner take all basis. Again, as trade exposure to China increases, there is “a net shift in favor of candidates on the right” (Autor, 2020, p. 3171). With respect to both ideological polarization and rightward shift in elections, the rightward movement is mostly a reflection of changes among non-Hispanic whites, with little to zero compensating effects among Hispanics and Blacks (Autor et al., 2020, p. 3142). Once again, the center (Democratic moderates) loses.

We said the effects are small but they are significant, not just statistically significant but politically significant, in the sense that the marginal increases in rightward movement may have shaped the electoral outcomes. The outcomes of both the 2016 and 2020 elections were decided by very close votes in a few states. As Autor et al. say (2020, p. 3175), “[…] a 50% ceteris paribus reduction in the China trade shock between 2000 and 2014 would have tipped the narrow Republican voter majority in the states of Pennsylvania, Wisconsin, and Michigan, leading to an Electoral College victory for Hillary Clinton, instead of a victory for Donald Trump.”

6. Conclusion

A paper about the USA–China trade, in today's political climate is bound to be of more than academic interest. We do not want to contribute to the frenzy that characterizes some parts of the USA media that focus only on the competitive aspects of the USA–China relations and forget that there are also gains on both sides in terms of consumer welfare, gains from trade and lower-interest rates available to corporations and individuals. However, the benefits of trade disproportionately focus on the consumption side and downplay the costs to producers, both capitalists and workers. There were indeed substantial costs to the USA workers. The costs to capitalists were less because firms reorganized and outsourced production.

USA–China trade has been politically salient due to a combination of factors. Increased trade with China took place rapidly (thus earning the label “China shock”) and the competition and job displacement took place in the manufacturing sector where “good jobs” were held by non-college graduates. Third, these jobs, while interspersed across several states, were overall geographically concentrated in coastal areas and the upper eastern heartland. The coastal areas did lose manufacturing jobs but they were replaced by high-end service jobs. Often this process of restructuring took place inside the same corporations that were downsizing and closing manufacturing operations (Bloom et al., pp. 1–3). Finally, the rightward movement in the American electorate was predominantly among non-Hispanic whites in close electoral districts. This affected electoral outcomes.

Much research still needs to be done. How much of the rightward shift in American politics is structural and hence ongoing? How much is due to durable features of capitalism, like automation and how much to more time-limited effects like the China trade, whose economic impacts on jobs may be temporally limited. Bloom and his coauthors, based on their own careful data analysis, state that the employment effects of China trade ended in 2007 (Bloom et al., 2019). If this is the case, then the political effects nevertheless linger.

We have taken on only a small slice of the question of trade, jobs and domestic politics. To resolve these issues in a satisfactory way, we need to put this narrow question into a much more general context. If jobs have shifted from the heartland and upper mid-west to the coasts, do we want to encourage “place-based policies” as Austin et al. do in “Jobs for the Heartland” (2018). Or believing that building human capital is the answer, do we want to advocate affordable college education? The decline of internal migration strengthens the case for place-based policies. Is the task of government to bring work to places where people live or encourage people to follow changing locations of economic activity?

When I began this paper, my hunch was that changes in the USA manufacturing jobs were the result of the long-term bumping and grinding of tectonic forces related to shifting comparative advantages. Indeed, these forces are important as the USA slowly shifts from technologies associated with the heavy industrial revolution, i.e. steel, cars, refrigerators etc. to a manufacturing and service economy where digitized information, electronics and artificial intelligence are central. In the long run, it is difficult to stem the force of shifting comparative advantages except by draconian protectionist measures. For good reasons, this is not the direction in which the world has been moving since the end of the Second World War or is this secular course likely to be reversed.

Notes

1. These are topics recently taken on by both political scientists and economists. In economics see Autor et al. (2020) on “Importing Political Polarization: The Electoral Consequences of Rising Trade Exposure”; Bloom et al. (2019) “The Impact of Chinese Trade on US Employment: the Good, the Bad, and the Debatable.” In political science, see Mutz and Mansfield (2021), “The Racialization of International Trade.” For a path-breaking account of the relationship of trade to political alignments, based on political extensions of the famous Stolper–Samuelson theorem, see Ronald Rogowski (1989), Commerce and Coalitions.

2. Many would argue that there is nothing unexpected or “wrong” about Chinese import competition displacing the USA employment, at least, to the extent that such displacement takes place on the basis of comparative cost considerations. But of course, it is precisely this point which is politically salient. Artificially low wages (e.g. low because the organization of labor is disallowed) subsidies to particular industries, favoritism to defense industries with large spillovers to connected industries, differential access to capital and rent-free land are examples of how trade can be distorted. Nor do we assume that all distortions take place in China and not in the USA.

3. For example, if the USA raised tariffs prohibitively against China but did not increase its overall rate of savings (i.e. if overall demand remained constant), some other country would take up the slack in demand. This would reallocate the overall USA deficit among different trading partners, but it would not reduce the totals. This is not to deny, harking back to Hirschman's National Power and the Structure of Foreign Trade (1945) that the distribution of the deficit (or surplus) among trade partners, might itself be a defensible goal of foreign economic policy. From the perspective of the USA, it makes a difference if the trade imbalance is with the United Kingdom vs China.

4. Indeed, even the abstract economic models fall short in that they downplay the potential for successful protectionist measures (tariffs, quotas and non-tariff barriers) on the part of holders of factors that are relatively scarce. If struggles for protection are successful, they will obviously blunt the income effects of trade liberalization. There are numerous examples of successful defensive measures on the part of rent-seeking coalitions in the face of liberalizing influences, e.g. Bismark's coalition of large landowners and industrialists (the famous coalition of iron and rye) against labor in the late 19th century, and the reaction of textile producers and clothing manufacturers in the 1960s and 1970s in the USA to cheaper imports from newly industrializing countries.

5. Technically, the Stolper–Samuelson theorem asks us to look at the supply of factors as ratios in two senses: first, the internal (domestic) ratios and second, a comparison of domestic ratios with those of other countries or the world at large. For example, what are the capital/labor ratios in the USA relative to the rest of the world or relative to a smaller unit useful for bilateral or regional analysis?

6. A quick search of sources that have cited or written about the Stolper–Samuelson theorem comes up with 11,300 results. While I did not code the articles according to whether they appeared in economic or political science journals, a superficial perusal suggests a huge dominance of economics journals.